The Therapeutic Alliance in the Treatment of Traumatized Youths

Relationship to Outcome and Dropout Across Rater Perspectives and

Therapeutic Interventions

Silje Mrup Ormhaug

Norwegian Centre for Violence and Traumatic Stress Studies (NKVTS), Oslo, Norway

Submitted for the degree of PhD at the

Department of Psychology, Faculty of Social Sciences, University of Oslo
2015

























6LOMH0UXS2UPKDXJ

Series of dissertations submitted to the
Faculty of Social Sciences, University of Oslo
No. 566
,661


$OOULJKWVUHVHUYHG1RSDUWRIWKLVSXEOLFDWLRQPD\EH
UHSURGXFHGRUWUDQVPLWWHGLQDQ\IRUPRUE\DQ\PHDQVZLWKRXWSHUPLVVLRQ








&RYHU+DQQH%DDGVJDDUG8WLJDUG
3ULQWSURGXFWLRQ-RKQ*ULHJ$6%HUJHQ

3URGXFHGLQFRRSHUDWLRQZLWK$NDGHPLNDSXEOLVKLQJ2VOR
7KHWKHVLVLVSURGXFHGE\$NDGHPLNDSXEOLVKLQJPHUHO\LQFRQQHFWLRQZLWKWKH
WKHVLVGHIHQFH.LQGO\GLUHFWDOOLQTXLULHVUHJDUGLQJWKHWKHVLVWRWKHFRS\ULJKW
KROGHURUWKHXQLWZKLFKJUDQWVWKHGRFWRUDWH


Table of Contents
Acknowledgements ................................................................................................................................. i
Summary ............................................................................................................................................... iii
List of Papers ......................................................................................................................................... v
Tables and Figures ................................................................................................................................ v
List of Abbreviations ............................................................................................................................ vi
1. Introduction ....................................................................................................................................... 1
1.1 The Therapeutic Alliance in the Treatment of Traumatized Youth: How Central Is It? ............... 1
1.2 Background: Prevalence and Consequences of Childhood Trauma ............................................ 3
1.2.1 PTSD. .................................................................................................................................... 4
1.2.2 Theoretical models of PTSD. ................................................................................................ 6
1.2.3. Complex trauma.................................................................................................................... 7
1.3 Therapeutic Interventions for Traumatized Youth ........................................................................ 7
1.3.1 TF-CBT. ................................................................................................................................ 8
1.3.2 Understanding how treatment leads to change. ..................................................................... 9
1.4 The Therapeutic Alliance ............................................................................................................ 10
1.4.1 The alliance-outcome relationship in youth therapies. ........................................................ 10
1.4.2 The content and measurement of the youth alliance............................................................ 11
1.4.3 Youth and therapist perspectives on the alliance. ................................................................ 12
1.4.4 The role of caregivers. ......................................................................................................... 13
1.4.5 Including the systemic context: interplay between youth and caregiver perspectives......... 14
1.5 Disentangling the Alliance-Outcome Relationship ..................................................................... 15
1.5.1 Challenges to the role of the alliance as an agent of therapeutic change. ............................ 15
1.5.2. Potential pathways from the alliance to outcome. .............................................................. 17
2. The Present Study ........................................................................................................................... 23
2.1. Aims and Research Questions .................................................................................................... 23
3. Materials and Methods ................................................................................................................... 23
3.1 Procedure: The Norwegian TF-CBT Study................................................................................. 23
3.2 Sample ......................................................................................................................................... 26
3.2.1. Youth sample. ..................................................................................................................... 26
3.2.2 Caregiver sample. ................................................................................................................ 28
3.2.3 Therapist sample. ................................................................................................................. 28
3.3 Treatment Conditions.................................................................................................................. 29
3.3.1 TF-CBT. .............................................................................................................................. 29
3.3.2 TAU. .................................................................................................................................... 29
3.4 Measures ..................................................................................................................................... 30
3.4.1 Youth alliance. ..................................................................................................................... 30
3.4.2 Caregiver alliance. ............................................................................................................... 30
3.4.3 Therapist alliance. ................................................................................................................ 31
3.4.4 Youths perceptions of parental approval of treatment........................................................ 31
3.4.5 Trauma exposure. ................................................................................................................ 32
3.4.6 Self-reported PTSS. ............................................................................................................. 32

3.4.7 Clinician-rated PTSS. .......................................................................................................... 32

3.4.8 Caregiver-rated PTSS. ......................................................................................................... 33
3.4.9 Depressive symptoms. ......................................................................................................... 33
3.4.10 Anxiety symptoms. ............................................................................................................ 33
3.4.11 General mental health. ....................................................................................................... 34
3.4.12 Youth-rated treatment satisfaction. .................................................................................... 34
3.5 Statistical Analyses ..................................................................................................................... 34
3.5.1 Initial analyses (papers I, II & III). ...................................................................................... 34
3.5.2 Hierarchical regression analyses (papers I & II).................................................................. 34
3.5.3 Exploratory Factor Analyses (paper II). .............................................................................. 35
3.5.4 Logistic regressions (paper III). ........................................................................................... 36
3.5.5 Handling missing data (papers I, II and III)......................................................................... 36
3.5.6 Statistical software. .............................................................................................................. 37
3.6 Ethical Considerations ................................................................................................................ 37
4. Results .............................................................................................................................................. 38
4.1 Paper I: The Therapeutic Alliance in Treatment of Traumatized Youth: Relation to Outcome in
a Randomized Clinical Trial ............................................................................................................. 38
4.2 Paper II: Therapist and Client Perspectives on the Alliance in the Treatment of Traumatized
Adolescents ....................................................................................................................................... 39
4.3 Paper III: Understanding Dropout in the Treatment of Traumatized Youths: Background,
Treatment, and First Session Process Variables .............................................................................. 40
5. Discussion ......................................................................................................................................... 40
5.1 Discussion of Main Findings ...................................................................................................... 41
5.1.1 The alliance is a significant predictor of treatment process and outcome. .......................... 41
5.1.2 Youths and therapists views of the alliance are not interchangeable. ............................... 43
5.1.3 Youths perceptions of caregiver approval of therapy predict dropout. .............................. 43
5.1.4 Linking the alliance to outcome: evaluating results in light of the pathways model. .......... 44
5.1.5 The good news: the conditions for the alliance may be better than assumed. ..................... 47
5.2 Methodological Considerations .................................................................................................. 48
5.2.1 The criterion validity of the therapeutic alliance scale. ....................................................... 48
5.2.2 The internal reliability of the questionnaires ....................................................................... 48
5.2.3 Timing and source of ratings. .............................................................................................. 49
5.2.4 Attrition and missing data. ................................................................................................... 50
5.2.5 The nested nature of the data. .............................................................................................. 51
5.2.6 The external validity of the findings. ................................................................................... 51
5.2.7 Interpretation of non-findings. ............................................................................................. 53
5.3 Clinical Implications ................................................................................................................... 53
5.4. Recommendations for Future Research ..................................................................................... 54
6. Conclusions ...................................................................................................................................... 56
References ............................................................................................................................................ 57
Papers I - III
Appendices
1a. CAPPATS, English version
1b. CAPPATS, Norwegian version

Acknowledgements
This thesis is part of the Norwegian TF-CBT study that was carried out at The
Norwegian Centre for Violence and Traumatic Stress Studies (NKVTS). This is a large
treatment study that started up in 2008, and many are those who deserve warm words of
thanks for their contributions to the TF-CBT study and this thesis. First of all, I would like to
thank all the youths and their caregivers who participated in the TF-CBT study. They shared
with us of their time and their experiences, and provided us with important knowledge that
can help increase our understanding of how clinicians can provide better treatment for
traumatized youths.
During my doctoral work I have received guidance from two excellent supervisors.
First, professor Tine K. Jensen has provided invaluable support to my work with this thesis
both as the project leader of the TF-CBT study and as my primary supervisor. She has
generously shared of her knowledge and ideas, her sharp eye, and her professional network,
and her encouragement and enthusiasm has been an important inspiration. Tine: You shall
know that your caring involvement in my work has been recognized, and that both my mum
and my husband speak very warmly of you! I must also thank my second supervisor,
Professor Helge Rnnestad. His extended knowledge of the field of psychotherapy research
and insightful comments have been very useful, and I am particularly grateful that he was able
to provide help in the last phase of this work, even though his calendar was already fully
booked.
In addition to my supervisors, I have also been so fortunate to collaborate with other
very skilled researchers. First I must thank Professor Stephen R. Shirk for his involvement in
this thesis both as a co-author and as an informal mentor. He generously invited me to stay
with his research group at the University of Denver, and the six months I spent at the Shirk
lab were exciting, inspiring and very educational. I must also thank the other members of the
lab JP, Patty, Tess, Emma, and Ryan for taking good care of me, and showing me all the
fun parts of Denver! Here in Oslo Tore Wentzel-Larsen (a.k.a. ToRe) has been an excellent
helper, co-author, and teacher in statistics. He has enabled me to understand and perform
analyses that seemed impossible at start, and although I am still far from a fluent speaker of
R, I do at least know some of the basics of that statistical universe now.
During my work I have been so lucky to be surrounded by great colleagues at
NKVTS. First and foremost I must thank Dr. Tonje who has been a great support during the
i

whole period. With her warmth, energy, and working capacity she made an enormous
contribution during the years the two of us spent recruiting participants and collecting data to
the TF-CBT study. Tonje: you shall know that all the serious discussions and the silly chitchats, the travels and the training sessions we have had throughout the years have meant a lot
to me! I also want to thank the rest of the TF-CBT group and our associated helpers: Shirley,
Karina, Live, Lene, and Tore I. The TF-CBT study and this thesis could not have been
conducted without them. Also, warm thanks to the director of NKVTS, Inger Elise, and the
leader of the Child and adolescent section, Mona-Iren, for supporting me and our study all the
way, and giving me the extra time I needed to finish this work.
Finally I would like to thank my friends and family that have been there for me during
these years. I must in particular thank my parents Gitte and Terje. They have always
supported my work, and have generously taken care of Jon (1.5), cooked dinners and helped
us with house chores during times of illness and upcoming deadlines. Mamma & pappa: your
practical and emotional support has been invaluable! Last, but not least, I want to thank my
dear friend and husband Hvard: With your loving care for Jon and me, your fresh baked
bread, and sometimes extremely silly jokes you have made me laugh also during the most
strenuous periods of work, and your help has been essential to enable me finish this thesis.
Together with Jon you remind me every day of what matters most in life!

Silje M. Ormhaug
Oslo, March, 2015

ii

Summary
Traumatizing events such as domestic violence, severe accidents, and sexual abuse
place youths at risk of developing mental health problems, and many will be in need of
therapy. To date, a variety of treatment models have been developed, and although many of
these models show promise in alleviating youths posttraumatic symptoms and related
psychological problems, little is known about the therapeutic change processes involved in
these treatments. One process variable that has been found to significantly predict outcome
across a variety of interventions and diagnostic disorders is the therapeutic alliance. Although
a strong alliance is also assumed to be pivotal in the treatment of traumatized youth, this
assumption has only been rarely tested empirically. The overarching aim of this thesis is to
better understand the relationship between the therapeutic alliance and dropout and outcome
in the treatment of traumatized youth. Data were derived from a Norwegian randomized
effectiveness trial comparing Trauma-Focused Cognitive Behavioral Therapy (TF-CBT) to
therapy as usual (TAU) in regular outpatient clinics. The sample consisted of 156 youth (M
age = 15.1 years; range: 1018 years) showing significant symptoms of posttraumatic stress
(PTSS), 91 caregivers, and 71 therapists. All participants reported on their perspectives on the
alliance (Jensen et al., 2014).
The first area of investigation involved the relationship between the therapeutic
alliance and outcome (paper I). The results showed that youth-rated alliance assessed midtreatment was a significant predictor of lower symptom levels post-treatment, and that this
relationship was moderated by treatment condition. Specifically, a strong alliance was
significantly related to better outcomes in the TF-CBT condition but not in TAU. This study
is one of the first to provide a direct comparison of the alliance-outcome relationship across
treatment models, and the findings indicate that there is an important interaction between the
alliance and the therapeutic approach. It seems that a positive working relationship is
especially important in the context of TF-CBT, which requires youth involvement in specific
therapy tasks but may be less related to change in the more unspecific TAU condition.
In the second paper, the therapists perspective on the alliance and its relationship to
youths evaluations and outcome was examined (paper II). In addition, the consequences of
discrepant youth and therapist ratings were investigated. The results showed that youth ratings
of the alliance were significantly related to outcome and treatment satisfaction. Therapist
ratings predicted youths treatment satisfaction, but were not related to post-treatment
iii

symptom reduction. Furthermore, associations between youth and therapist ratings were only
moderate, and analyses showed that the perspectives differed in their underlying factor
structure. Youth seem to separate the alliance into a positive and negative dimension;
therapists ratings cluster into the theoretical dimensions task and bond. Level of alliance
agreement was related to outcome, and dyads where the youth reported poorer alliances
compared with the therapist were associated with higher residual PTSS and lower treatment
satisfaction. These findings imply that youth and therapist perspectives are not
interchangeable, and that therapists cannot assume that the youth share their views of the
alliance. The results suggest that therapists should investigate directly how youth perceive the
alliance, since the youth-rated alliance is an important predictor of outcome.
The last aim of this thesis was to learn more about the process variables involved in
treatment attendance (paper III). Specifically, first session alliance ratings from youth,
therapist and caregiver alliances were assessed as predictors of dropout, in addition to youth
background variables and treatment-specific factors. Furthermore, a new scale was developed
(the Child- and Adolescent-Perceived Parental Approval of Therapy, CAPPATS) to evaluate
to what degree the youth perceived that their parents supported the treatment, and whether
this perceived support was related to dropout. The results showed that dropout was predicted
by therapist-rated alliance, youths perceptions of caregiver approval of therapy and a lack of
caregiver participation, but not by youth background characteristics or the treatment method.
The findings indicate that more attention should be paid to the in-treatment process variables
in order to gain a better understanding of which youth are at risk of dropping out. Specifically,
further investigations of the role of caregivers involvement in the treatment and youths
perceptions of parental support seem warranted.
Collectively, the findings in this thesis contribute to a better understanding of the role
of the therapeutic alliance in the treatment of traumatized youth, how the alliance interacts
with treatment method to enact change, and how different raters perspectives are related to
outcome and process.

iv

List of Papers
,

Ormhaug, S. M., Jensen, T. K., Wentzel-Larsen, T. & Shirk, S. R. (2014). The

therapeutic alliance in treatment of traumatized youth: Relation to outcome in a
randomized controlled trial. Journal of Consulting and Clinical Psychology, 82(1), 52 doi:10.1037/a0033884

,,

Ormhaug, S. M., Shirk, S. R., & Wentzel-Larsen, T. Therapist and Client Perspectives
on The Alliance in the Treatment of Traumatized Adolescents (manuscript submitted
for publication)

,,,

Ormhaug, S. M., & Jensen, T. K. Understanding Dropout in the Treatment of

Traumatized Youths: Background, Treatment, and First Session Process Variables
(manuscript submitted for publication)

Tables and Figures

Table 1: Short overview of the main focus and different subsamples in each paper (p. 26)
Table 2: Participant characteristics (p. 27)
Figure 1: Proposed model of potential pathways from alliance to outcome (p. 19)
Figure 2: Participant flowchart (p. 25)
Box 1: Diagnostic criteria for posttraumatic stress disorder (p. 5)

A slightly revised version of this paper has been published in European Journal of Psychotraumatology,

6:

List of Abbreviations

Chronbachs alpha

CAPPATS

Child- and Adolescent-Perceived Parental Approval of Therapy Scale

CAPS-CA

Clinician-Administered PTSD Scale for Children and Adolescents

CPSS

Child PTSD Symptom Scale

DSM-IV

Diagnostic and Statistical Manual of Mental Disorders, 4th edition

DSM-5

Diagnostic and Statistical Manual of Mental Disorders, 5th edition

DV

Dependent variable

EFA

Exploratory factor analyses

IV

Independent variable

NKVTS

Norwegian Centre for Violence and Traumatic Stress Studies

MFQ

Mood and Feelings Questionnaire

Mplus

Statistical software

PTSS

Posttraumatic stress symptoms

PTSD

Posttraumatic stress disorder

Pearsons correlation coefficient

Statistical software

SCARED

Screen for Child Anxiety-Related Disorders

SD

Standard deviation

SDQ

Strength and Difficulties Questionnaire

SPSS

Statistical Package for the Social Sciences (statistical software)

TASC-r

Therapeutic Alliance Scale for Children- revised

TAU

Treatment as usual

TF-CBT

Trauma-focused cognitive behavioral therapy

Chi squared

vi

1. Introduction
1.1 The Therapeutic Alliance in the Treatment of Traumatized Youth: How Central Is
It?
It is commonly assumed that a strong alliance is essential for the successful treatment
of traumatized youth (see e.g., J. A. Cohen, Mannarino, Kliethermes, & Murray, 2012;
Kearney, Wechsler, Kaur, & Lemos-Miller, 2010; Lawson, 2009; Shirk & Eltz, 1998).
However, prospective studies of this relationship variable are rare. Most studies investigating
the benefits of treatments for traumatized youth have focused on therapeutic techniques and
comparisons between different treatment approaches, instead of the therapeutic alliance and
other relational aspects of these treatments. This fact is problematic because a better
understanding of the relational context of the therapeutic interventions will help increase our
understanding of how and why psychotherapy works (Kazdin, 2009; Norcross & Lambert,
2011; Orlinsky, Rnnestad, & Willutzki, 2004), and may improve the implementation of
effective treatments into regular clinics (Kazdin & Nock, 2003).
One exception is an early study by Eltz and colleagues (Eltz, Shirk, & Sarlin, 1995);
these authors found that a strong therapeutic alliance was significantly related to better
progress in the treatment of maltreated adolescents. Furthermore, a growing body of studies
on adult patients has found that the therapeutic alliance is a significant predictor of outcome
in treatments of posttraumatic stress disorder (PTSD) (Cloitre, Koenen, Cohen, & Han, 2002;
Cloitre, Stovall-McClough, Miranda, & Chemtob, 2004; Keller, Zoellner, & Feeny, 2010;
McLaughlin, Keller, Feeny, Youngstrom, & Zoellner, 2013). Although these findings from
adult studies are important and can help inform the field of youth trauma, developmental
aspects are likely to influence the presentation and treatment of youths post-traumatic
reactions and results from adult studies cannot be directly transferred to the treatment of
PTSD for youth. For example, youth may lack an understanding of how psychological
problems develop and what it may take to solve them, and they may find it difficult to see the
link between the tasks performed in treatment and the subsequent reduction of symptoms
(Shirk & Saiz, 1992). These developmental aspects may again influence the relative
importance of the therapeutic alliance as a mutual and collaborative phenomenon.
Furthermore, since children and adolescents are dependent on their caregivers, caregivers are
often involved in the therapies. This fact implies that there are several alliances to be
negotiated, and that the child-therapist alliance, caregiver-therapist alliance and the
1

relationship between the child and caregiver alliances may substantially influence the
treatment process (Zack, Castonguay, & Boswell, 2007). As a result, the therapeutic alliance
has to be studied within a child and adolescent treatment context if we are to better understand
the role it plays in the process and outcome of youth trauma treatments. The main aim of this
thesis is to address current knowledge gaps in the youth trauma field and investigate the role
of the therapeutic alliance and related relationship variables in youth trauma treatments. In
particular, a primary goal of this study is to learn more about the predictive value of the
therapeutic alliance in relation to symptom reduction (papers I and II), treatment satisfaction
(paper II) and dropout (paper III).
Furthermore, it is an open question whether the alliance-outcome relationship is the
same across different treatment conditions. Based on meta-analyses of adult studies, there are
reasons to believe that the alliance is a consistent predictor of outcome across treatment
models (Flckiger, Del Re, Wampold, Symonds, & Horvath, 2012). On the other hand, direct
comparisons between different treatment conditions indicate that the alliance may play a
different role in different treatment conditions, both in adult (Arnow et al., 2013; Ulvenes et
al., 2012) and adolescent therapy (Cummings et al., 2013; Hogue, Dauber, Stambaugh,
Cecero, & Liddle, 2006). This finding implies that there might be an interaction between the
alliance and the type of treatment provided. This question was investigated in the first paper
in this thesis, where the predictive value of youth-reported alliance was compared across two
different treatment conditions.
Another question is related to the therapists perspectives on the alliance, and how
these are related to youths alliance evaluations. Therapists perspectives on the alliance are
important because it is the therapist that is responsible for managing the therapeutic process
(Safran, Muran, & Eubanks-Carter, 2011), and his or her evaluations of the alliance are likely
to influence in-session decision-making about specific interventions. Additionally, since the
alliance is an interpersonal construct, failure to recognize the youths perspective on the
alliance, as reflected in discrepant alliance ratings, could indicate a lack of therapist
attunement to the youths experience and predict poorer outcome. Thus far, at least one study
has found that a failure to recognize and repair ruptures in the alliance predicted worse
outcome in the treatment of adults with PTSD (McLaughlin et al., 2013). Better
understanding the relationship between youth and therapist ratings and investigating potential
sources of this divergence is the aim of the second paper in this thesis.
Several studies have shown that dropout is a common problem in youth trauma
treatments (Gopalan et al., 2010; Lau & Weisz, 2003). So far, the majority of studies of
2

dropout in youth PTSD treatments have investigated youth background and demographic
variables; less attention has been paid to the in-treatment process variables (Chasson, Vincent,
& Harris, 2008; Eslinger, Sprang, & Otis, 2012; Sprang et al., 2013). In particular, both
youth-therapist and caregiver-therapist alliances hold promise as important predictors of
dropout. In addition, both theory and empirical studies suggest that there may be an important
relationship between youths and caregivers perceptions of the treatment that significantly
influences youths treatment attendance (Jensen et al., 2010; Robbins, Turner, Alexander, &
Perez, 2003). These aspects are explored as predictors of treatment dropout in the third paper
of this thesis.
There is currently little knowledge of how the alliance is related to outcome, i.e., the
mechanisms through which the alliance may instigate change. The final aim of this thesis is to
integrate the findings from the three studies discussed above in order to see whether these
findings can enhance our understanding of the pathways from the alliance to dropout and
outcome in the treatment of traumatized youth.
Data were derived from a randomized clinical trial investigating the process and effect
of trauma treatment in Norwegian community clinics. In this study, Trauma-Focused
Cognitive Behavioral Therapy (TF-CBT) was implemented in eight child and adolescent
clinics and the effectiveness of TF-CBT was compared to therapy as usual (TAU) (Jensen et
al., 2014). The results showed that, on average, youth reported significant reductions in their
PTSD symptoms from pre- to post-treatment in both conditions. However, youth in the TFCBT condition reported significantly larger improvement on a variety of symptom measures
compared with the youth in the TAU group (mean Cohens d = 0.47). Furthermore, in both
conditions, approximately 20% of the youth dropped out from the treatment and did not
benefit from the therapy provided. Better understanding the role the therapeutic alliance plays
in the treatment process and outcome will help inform therapists and contribute to better
caring for trauma-exposed children and adolescents.

1.2 Background: Prevalence and Consequences of Childhood Trauma

Every year, a substantial number of youth are exposed to potentially traumatic events
such as severe accidents, domestic violence, physical and sexual abuse, natural disasters, and
war (Copeland, Keeler, Angold, & Costello, 2007; Finkelhor, Ormrod, & Turner, 2007). In
Norway, there are currently no data on the prevalence of youth exposed to all types of
traumatic events. However, three recent studies have found that the number of youth affected
by violence, abuse, and sexual assault is high. In a sample of 7,033 high school seniors, 25%

of the students reported exposure to at least one episode of physical abuse from their
caregivers, and 15% of the girls and 7% of the boys reported they had been exposed to sexual
assault at some point in their life (Mossige & Stefansen, 2007). In another study of 15,930
10th graders, 23.6% of the boys and 11.8% of the girls reported exposure to peer violence
during the last year (Schou, Dyb, & Graff-Iversen 2007). The most recent study, in which a
representative sample of adults was asked about their lifetime exposure to violence and abuse,
33.6% of the women and 11.3% of the men reported they had experienced some kind of
sexual assault before the age of 13 (Thoresen & Hjemdal, 2014). All of these events put
children and adolescents at risk of developing severe mental health problems such as PTSD,
anxiety, depression, conduct disorders, social problems, substance abuse and school-related
problems (Dube, Felitti, Dong, Giles, & Anda, 2003; Gerson & Rappaport, 2013; Kilpatrick
et al., 2003). According to a recently published meta-analysis, the average rate of PTSD
among trauma-exposed children and adolescents is 15.9%. However, the study showed that
the prevalence varied according to the type of trauma and gender; girls exposed to
interpersonal trauma were at the highest risk (32.9%) (Alisic et al., 2014). Consequently, there
are reasons to expect that the prevalence of traumatized youth is even higher in clinical
settings. Although no national data exist, at least two studies conducted in Norwegian child
and adolescent mental health clinics support this assumption. The first study was carried out
as part of the Norwegian TF-CBT trial. Here, the referred youth were screened for trauma at
intake and the results showed that 47% of the youth reported exposure to at least one
traumatizing event (Ormhaug, Jensen, Hukkelberg, Holt, & Egeland, 2012). The other study
found that 60% of the youth in treatment reported exposure to child abuse (Reigstad,
Jrgensen, & Wichstrm, 2006). These figures are in line with international studies (Lau &
Weisz, 2003; McKay, Lynn, & Bannon, 2005). Of the trauma-exposed youth, between 42%
and 90% have been found to report significant levels of posttraumatic stress symptoms
(PTSS) (Kearney et al., 2010; Ormhaug et al., 2012). If these trauma reactions remain
unresolved, they are likely to have a lifelong, negative impact on psychological and social
well-being (Anda et al., 2006; Dube et al., 2003; McGloin & Widom, 2001). Consequently, it
is important that therapists in regular clinics know how to provide the best help for these
affected youth in order to alleviate their post-trauma reactions and regain a normal
developmental path.
1.2.1 PTSD. The primary aim of the Norwegian TF-CBT trial was to learn more about
effective therapy for youth suffering from PTSD and PTSS. PTSD was defined according to
the Diagnostic and Statistical Manual of Mental Disorders, 4th edition text revision (DSM-IV4

TR) criteria (APA, 2000; Box 1). In 2013, a new diagnostic manual was released, the DSM-5,
which included some changes to the PTSD diagnosis (APA, 2013; Box 1).
Box 1. PTSD
DSM-IV-TR criteria for PTSD (APA, 2000)
Criterion A: Exposure
The person has been exposed to a traumatic event in which both of the following have been present:
(1) the person has experienced, witnessed, or been confronted with an event or events that involved actual or threatened
death or serious injury, or a threat to the physical integrity of self or others
(2) the persons response involved intense fear, helplessness, or horror. Note: In children, this may be expressed instead
by disorganized or agitated behavior
Criterion B: Intrusion
The traumatic event is persistently re-experienced one (or more) of the following ways:
(1) recurrent and intrusive distressing recollections of the event, including images, thoughts, or perceptions. Note: In
young children, repetitive play may occur in which themes or aspects of the trauma are expressed
(2) recurrent distressing dreams of the event. Note: In children, there may be frightening dreams without recognizable
content.
(3) acting or feeling as if the traumatic event were recurring (includes a sense of reliving the experience, illusions,
hallucinations, and dissociative flashback episodes, including those that occur upon awakening or when intoxicated).
Note: In young children, trauma-specific re-enactment may occur.
(4) intense psychological distress at exposure to internal or external cues that symbolize or resemble an aspect of the
traumatic event.
(5) physiologic reactivity upon exposure to internal or external cues that symbolize or resemble an aspect of the
traumatic event.
Criterion C: Avoidance
Persistent avoidance of stimuli associated with the trauma and numbing of general responsiveness (not present before the
trauma), as indicated by three (or more) of the following:
(1) efforts to avoid thoughts, feelings, or conversations associated with the trauma
(2) efforts to avoid activities, places, or people that arouse recollections of the trayma
(3) inability to recall an important aspect of the trauma
(4) markedly diminished interest or participation in significant activities
(5) feeling of detachment or estrangement from others
(6) restricted range of affect (e.g., unable to have loving feelings)
(7) sense of a foreshortened future (e.g., does not expect to have a career, marriage, children, or a normal life span)
Criterion D: Arousal
Persistent symptoms of increased arousal (not present before the trauma), as indicated by two (or more) of the following:
(1) difficulty falling or staying asleep
(2) irritability or outbursts of anger
(3) difficulty concentrating
(4) hypervigilance
(5) exaggerated startle response
Criterion E: Duration
Duration of the disturbance (symptoms in Criteria B, C, and D) is more than 1 month.
Criterion F: Impaired functioning
The disturbance causes clinically significant distress or impairment in social, occupational, or other important areas of
functioning.
Alterations to the PTSD diagnosis in DSM-5 (APA, 2013)
- The A2 criterion is removed
- Symptoms are clustered into four instead of three factors:
B. Re-experiencing
C. Avoidance
D. Altered and persistent negative mood and cognitions
E. Arousal

Thus far, there are reasons to assume that the alterations to the PTSD diagnosis will not have
major effects on clinical fields, since studies have indicated that the DSM-5 criteria will lead
to similar PTSD rates in both adolescents (Hafstad, Dyb, Jensen, Steinberg, & Pynoos, 2014)
and adults (Elhai et al., 2012; Kilpatrick et al., 2013).
1.2.2 Theoretical models of PTSD. In order to help youth overcome their PTSS,
understanding more of how the symptoms develop and are maintained is important. One
central theory that has helped inform several treatment methods is the cognitive model
proposed by Ehlers and Clark (2000). In this model, it is suggested that PTSS become
persistent if the traumatic event is processed and stored in memory in a way that makes the
person feel that the situation is a current ongoing threat rather than a time-limited event. This
sensation is assumed to be the result of 1) excessively negative thoughts and appraisals of the
trauma and/or its consequences, and 2) a lack of sufficient elaboration and contextualization
of the event in the persons autobiographical memory, combined with strong perceptual and
associative priming. Taken together, these cognitive processes will make affected individuals
feel that they are still in danger and that the traumatic event has global and negative
consequences for their future. It is furthermore suggested that these maladaptive assumptions
are maintained by a series of problematic behavioral and cognitive strategies such as safety
behaviors, avoidance and selective attention to threat cues. Studies of traumatized youth have
supported this theory, and found that in particular maladaptive appraisals are involved in the
development and maintenance PTSS over time (Meiser-Stedman, Dalgleish, Glucksman,
Yule, & Smith, 2009; Stallard & Smith, 2007). In order to alleviate PTSS, this theory implies
that it is important to help the child reprocess the trauma and develop a more coherent
memory of the trauma. It is also important to promote behaviors that can help extinguish the
link between trauma reminders and the sense of fear and ongoing threat so that the traumatic
event is no longer perceived as a current danger. In therapy, the youths fear reactions related
to the traumatic event may make it challenging to complete the exposure tasks. One can thus
expect that a strong therapeutic alliance is significantly related to youths ability to remain in
treatment and complete these tasks.
From a developmental perspective, there are a variety of factors influencing a childs
post-trauma adjustment. One framework to understand these factors and how they influence
each other is the developmental psychopathology model of childhood traumatic stress
proposed by Pynoos, Steinberg, and Piacentini (1999). In this model, intrinsic child factors
such as cognitive maturity and developmental level are important, in addition to contextual

factors such as caregiver functioning, social support, trauma reminders and secondary
stressors following the trauma. Together, these factors influence the childs ability to
contextualize and make sense of the traumatic event, to avoid the development of excessive
negative appraisals and promote helpful behaviors. Similarly, a recent meta-analysis
investigating risk factors for the development of PTSD found that in addition to the childs
subjective experience of the traumatic event, post-trauma variables such as family functioning
and social support were the strongest predictors of PTSD (Trickey, Siddaway, MeiserStedman, Serpell, & Field, 2012). This finding implies that in treatment, in addition to
working with the individual child and his or her traumatic memories and post-trauma
cognitions, therapist should involve the caregivers and focus on how the childs support
system can contribute in the childs recovery.
1.2.3. Complex trauma. In the Norwegian TF-CBT study, a substantial fraction of the
participating youth (59%) had been exposed to multiple and chronic traumas within the
caregiving relationship. These events include domestic violence, child maltreatment and intrafamilial sexual abuse; studies have found that such events are associated with an increased
risk of developing a series of relational and behavioral problems that are often referred to as
complex trauma (J. A. Cohen et al., 2012; Cook et al., 2005). In particular, interpersonal
problems are often prominent. According to attachment theory (Bowlby, 1988), a child will
develop expectations about interactions and relationships with other people based on his or
her early experiences with his or her caretaker(s). These expectations form inner working
models that will guide the child in his/her interactions with other people later in life, such as
peers, teachers and romantic partners. If the attachment relationship has not been safe,
engaging in a new relationship may be a trauma trigger for the youth and lead to increased
levels of vigilance and mistrust. For instance, several studies have found that persons exposed
to early trauma have relationship problems and automatic harm assumptions (Cloitre, Cohen,
& Scarvalone, 2002; DePrince, Combs, & Shanahan, 2009; Furman, Simon, Shaffer, &
Bouchery, 2002). Because the therapist-child relationship bears much resemblance to the
caregiver-child relationship, establishing a strong therapeutic alliance with youth with
complex trauma may be particularly challenging (J. A. Cohen et al., 2012; Eltz et al., 1995;
Shirk & Eltz, 1998).

1.3 Therapeutic Interventions for Traumatized Youth

During the last decade, there have been an emerging number of trials investigating the
effects of psychological treatments for children and adolescents suffering from PTSD and

PTSS (Carr, 2004; Silverman, Ortiz, & Viswesvaran, 2008). These studies include a variety of
treatment types such as exposure-based cognitive behavioral treatments (CBTs), Eye
Movement Desensitization and Reprocessing (EMDR) therapy, client-centered therapy,
family therapy and different forms of group therapy. Although many of these models report
widespread support in terms of clinical experience and client satisfaction, the empirical
evidence for their efficacy is less clear. In a literature review by Silverman and colleagues
(2008), 21 studies of different treatment models were classified along a continuum of
methodological rigor. According to the authors, the majority of the treatment models were
classified as possibly efficacious or experimental. One treatment model was classified as
probably efficacious (School-Based Group CBT), and it was only TF-CBT that met the wellestablished criteria.
1.3.1 TF-CBT. TF-CBT is a short-term, component-based intervention developed by
J. A. Cohen, Mannarino, and Deblinger (2006). The model builds on elements from cognitive,
behavioral, interpersonal and family therapy, in addition to trauma theory. It involves work
with the youth and their caregivers in both parallel and conjoint sessions. TF-CBT is normally
provided over a course of 1215 sessions. However, for youth exposed to more severe and
complex trauma, the treatment is often expanded up to 25 sessions (J. A. Cohen et al., 2012).
The components included in the treatment are psychoeducation, teaching relaxation and
affective modulations skills, learning cognitive coping skills, working through the trauma
narrative, cognitive processing, in vivo mastery of trauma reminders, and enhancing safety
and future development. In addition, there is a focus on parenting skills throughout the
treatment (J. A. Cohen et al., 2006). To date, 14 randomized, controlled trials have been
published, all documenting lower levels of PTSS and other trauma-related symptoms in
participants receiving TF-CBT compared with the control condition. The studies include
children and adolescents exposed to a variety of traumatic events such as sexual abuse (J. A.
Cohen, Deblinger, Mannarino, & Steer, 2004; J. A. Cohen & Mannarino, 1996, 1998; J. A.
Cohen, Mannarino, & Kundsen, 2005; J. A. Cohen, Mannarino, Perel, & Staron, 2007;
Deblinger, Lippman, & Steer, 1996; Deblinger, Mannarino, Cohen, Runyon, & Steer, 2011;
Deblinger, Stauffer, & Steer, 2001; King et al., 2000), domestic violence (J. A. Cohen,
Mannarino, & Iyengar, 2011), natural disasters (Jaycox et al., 2010), war exposure and sexual
abuse (O'Callaghan, McMullen, Shannon, & Rafferty, 2013), and in mixed trauma samples
(Jensen et al., 2014; Scheeringa, Weems, Cohen, Amaya-Jackson, & Guthrie, 2011). In the
Norwegian TF-CBT study, the results showed that the model is effective also in regular

clinics with ordinary therapists who are not trained in advance to be experts in trauma (Jensen
et al., 2014).
1.3.2 Understanding how treatment leads to change. Although the studies discussed
above represent an important first step in order to provide better help to trauma-exposed
youth, less is known about the change processes involved in the treatment. So far, only two
studies have investigated the sources of symptom improvement in TF-CBT. The first study
was a dismantling study that aimed to understand how treatment length and the trauma
narrative component were related to outcome (Deblinger et al., 2011). The results showed that
youth receiving eight sessions of TF-CBT including the trauma narrative reported
significantly lower levels of abuse-related fear and general anxiety compared with youth
receiving 16 sessions of TF-CBT but no narrative work. This finding supports the assumption
that the trauma narrative is an important contributor of change. The other study investigated
the relationship between a caregivers emotional reactions and changes in his or her childrens
symptom levels (Holt, Jensen, & Wentzel-Larsen, 2014). The authors found that caregivers
experienced reductions in their own levels of distress during their childs treatment. This
reduction mediated changes in their childs depressive symptoms, but was not related to a
reduction in the childs PTSS. Both of these studies focused on specific treatment components
(i.e., the trauma narrative and parent work), and did not investigate the relational context in
which these tasks were provided. Little is therefore known about the relational context of
these components. For example, it could be that the effectiveness of the trauma narrative as a
change mechanism depends on the degree to which youth and therapists collaborate on this
task. It could also be that for a caregiver, a strong relational bond with a therapist may be
beneficial in and of itself and reduce feelings of hopelessness and distress, which can in turn
have a positive effect on the childs depressive feelings.
In studies of adults, there is a debate regarding the relative importance of the common
relationship variables versus specific treatment components in the treatment of PTSD. The
results from at least one meta-analysis showed that different treatment interventions were
equally beneficial, lending support to the relative importance of common factors over specific
techniques (Benish, Imel, & Wampold, 2008; Wampold et al., 2010). Other authors have
argued that there is evidence to suggest that a component such as exposure is pivotal in the
treatment of PTSD (Ehlers et al., 2010) and that at least for some types of disorders specific
techniques may play an important role (Marcus, O'Connell, Norris, & Sawaqdeh, 2014).
Although it is never a question of either-or, since all treatments include both common and
specific factors (Safran & Muran, 2000; Wampold & Budge, 2012), it will be useful for
9

therapists to know whether some interventions are more helpful than others, or if it is the
relational context of the therapeutic methods, such as the therapeutic alliance, that is more
important.

1.4 The Therapeutic Alliance

Several meta-analyses have found that the therapeutic alliance is a significant
predictor of outcome with average correlations ranging from r = .22 (Flckiger et al., 2012;
Martin, Garske, & Davis, 2000) to r = .28 (Horvath, Del Re, Flckiger, & Symonds, 2011).
To date, various terms have been used to describe this therapeutic relationship, such as the
helping alliance, therapeutic bond, working alliance, and therapeutic alliance. Although the
conceptualizations and measures are diverse (Elvins & Green, 2008; Martin et al., 2000), the
most commonly used definition is the tripartite model of the working alliance proposed by
Edward Bordin (1979). He suggested that the alliance is a trans-theoretical construct relevant
to all types of therapy and that it consists of three different but related dimensions: 1) an
emotional bond between the therapist and client; 2) agreement on the goals of the therapy;
and 3) agreement on the therapeutic tasks to be conducted. Central to this definition is that the
working alliance refers to the collaborative aspect of the therapeutic relationship. In this
thesis, the term therapeutic alliance will be used to capture this collaborative aspect of the
relationship.
1.4.1 The alliance-outcome relationship in youth therapies. In studies of children
and adolescents, the relationship between the therapeutic alliance and outcome is less clear
compared with in the adult field. Two recent meta-analyses report average correlations
between alliance and outcome of r = .14 (McLeod, 2011) and r = .22 (Shirk, Karver, &
Brown, 2011). These studies differ in scope since Shirk and his colleagues limited their
sample to studies of individual therapy in order to make comparisons with the adult alliance
literature whereas McLeod included a wider range of treatment modalities (e.g., group and
family therapy). Although McLeod (2011) reports that the inclusion of a broader range of
studies did not influence the overall result, there are reasons to assume that the discrepant
results are, at least in part, related to methodological constraints in the youth psychotherapy
research field (Elvins & Green, 2008; Shirk et al., 2011). For one, the number of prospective
studies with explicit measures of the therapeutic alliance is small. Whereas the latest adult
meta-analysis included 201 studies (Flckiger et al., 2012), only 16 studies met similar
inclusion criteria in the Shirk et al. (2011) analysis, and only 38 studies were included in the
McLeod (2011) study. This limited empirical base increases the risk of potential bias.

10

Furthermore, both meta-analyses found that variations between individual studies were large
(r range: .38 .53), and that the relationship between alliance and outcome was significantly
moderated by aspects such as youth age, type of disorder, treatment mode, source, and timing
of the alliance assessments. In addition, there is still a lack of consensus on how to best
conceptualize the therapeutic alliance in youth therapy (Green, 2006; Shirk et al., 2011; Zack
et al., 2007). Developmental aspects such as childrens implicit understanding of the alliance,
the inclusion of caregivers in treatment and childrens limited position to negotiate about the
goals and tasks of the treatment make it unclear to what degree Bordins model is suitable
(Green, 2006; Jensen et al., 2010; Shirk & Saiz, 1992). This fact indicates that although there
is a need for more studies investigating the link between alliance and outcome in the youth
field, additional aspects of the alliance should also be explored. These include investigations
of the relationship between alliance and treatment method, the content and dimensionality of
the alliance, and the associations between various rater perspectives.
In this thesis, the associations between the alliance and outcome were investigated
both from a youth perspective (paper I) and a therapist perspective (paper II). Comparisons
across treatment methods were made in paper I.
1.4.2 The content and measurement of the youth alliance. The lack of consensus
about the dimensions of the alliance in youth therapy is reflected in the wide variety of
different alliance measures that are used. In the meta-analysis by Shirk et al. (2011), the
authors found that the 16 studies had used 10 different scales. This diversity complicates the
comparison of findings, since each scale measures a slightly different construct, and it has
been argued that progress in the youth alliance research field hinges on the development of
alliance scales with known underlying factors (Elvins & Green, 2008).
In this thesis, the alliance was measured with the Therapeutic Alliance Scale for
Children-Revised (TASC-r, Shirk & Karver, 2010; Shirk & Saiz, 1992). The TASC-r was the
first scale developed specifically for a younger client population (Elvins & Green, 2008) and
is by now one of the most widely used child and adolescence alliance scales. Items in the
scale were developed to correspond to the bond and task collaboration dimensions. The goal
dimension was not included, partly because it was assumed that it would be difficult for a
child to understand the links between the tasks in therapy and the subsequent goals.
Similarly, community clinicians reported rarely discussing explicit goals with children (Shirk
& Saiz, 1992). Over the years, several studies have confirmed that the TASC-r has sound
psychometric properties (Creed & Kendall, 2005; Fjermestad et al., 2012; Kendall et al.,
2009; Langer, McLeod, & Weisz, 2011), and it has been shown to significantly predict
11

treatment outcome in multiple studies (Accurso, Hawley, & Garland, 2013; Kazdin,
Marciano, & Whitley, 2005; Kazdin, Whitley, & Marciano, 2006).
Although the TASC-r is based on a two-factor task bond model, this theorized model
has yet to be confirmed empirically. So far, only one study has investigated the underlying
factor structure of youth and caregiver ratings of the TASC-r (Accurso et al., 2013). The
authors employed multilevel exploratory factor analyses (EFA) and the results showed that a
one-factor between-informants and a two-factor within-informants solution showed the best
fit. However, the two factors were divided by item valence (i.e., whether the items were
positively or negatively worded), rather than by the theorized task and bond items. This result
corresponds to findings from factor analyses of a variety of youth alliance scales: Although
the majority of scales have been based on the three-dimensional model of Bordin (1979), all
but one study (Johnson, Ketring, & Anderson, 2013) have failed to replicate this model with
empirical data (see e.g., DiGiuseppe, Linscott, & Jilton, 1996; Faw, Hogue, Johnson,
Diamond, & Liddle, 2005; Fjermestad et al., 2012; Hogue et al., 2006).
Another question relates to the therapists perspective of the alliance and the factor
structure of the therapist scale of the TASC-r has yet to be investigated. In the adult field, it
has been found that clients and therapists conceptual understandings of the alliance only
partially overlap (Bachelor, 2013), and there is evidence to suggest that the same may also be
true in the youth field. In a study by DiGiuseppe et al. (1996), youth and therapist ratings of
the Adolescent Working Alliance Inventory (AWAI) were analyzed. The results showed that
while therapists viewed the alliance in accordance with Bordins theoretical model, youths
responses clustered into one single dimension. Learning more about therapists and youths
implicit views of the alliance by investigating the factor structure of the youth and therapist
versions of the TASC-r was the second aim of paper II.
1.4.3 Youth and therapist perspectives on the alliance. Therapists perspectives of
the alliance have so far received little attention, although this perspective is also important.
For one, therapist judgments about alliance strength are likely to influence their in-session
decision-making about the use of specific interventions. For example, the introduction of a
potentially challenging task such as exposure may depend on whether a therapist perceives
the alliance to be fragile or strong. Secondly, since the alliance is an intersubjective construct,
the degree of discrepancy between youth and therapist ratings may matter for outcome. Low
levels of agreement may reflect a lack of attunement between therapist and youth. In
particular, if therapists rate the alliance as being stronger than the youth do, this could indicate
that therapists are unaware of youths negative feelings toward them or the therapeutic task,
12

making it unlikely that they will adjust their treatment strategies to meet the youths
expectations and preferred tasks.
In the McLeod meta-analysis the author found that the associations between youth and
therapist ratings and outcome were similar, indicating that the two individual perspectives
may be equally important. On the other hand, several studies have shown that the level of
youth and therapist agreement on the alliance is on average small to moderate with
correlations ranging from r .23 .38 (Creed & Kendall, 2005; Eltz et al., 1995; Fjermestad et
al., 2012; Hawley & Garland, 2008; Kendall et al., 2009; Shirk, Gudmundsen, Kaplinski, &
McMakin, 2008). This resembles findings in adult studies where the average agreement
between therapist and client rated alliance has been reported to be .36 (Tryon, Blackwell, &
Hammel, 2007). This fact means that although there is a certain degree of convergence
between youth and therapist perspectives, there seems also to be important differences
between youths and therapists perceptions of the alliance. These findings suggest that
therapists can only partially assume that their own perspectives of the alliance are shared by
the youth. Learning more about how youth and therapist ratings are related and whether the
degree of discrepancy had implications for outcome is the third aim of paper II.
1.4.4 The role of caregivers. Caregivers can be involved in their childs treatment in
different ways. At the minimum level, caregivers are often the ones initiating the referral, and
they are commonly in charge of scheduling appointments, transportation and other practical
issues related to the treatment process (Nock & Ferriter, 2005; Nock & Kazdin, 2001). As a
result, the caregivers are the gatekeepers of treatment and important agents of the childs
treatment attendance. Recently, there has been an increased awareness of the importance of
including the caregiver-perspective on the alliance in studies. In the meta-analysis by McLeod
(2011), the caregiver perspective was included in almost half of the studies (17 of the 38
studies); in the Shirk et al. (2011) analysis, in which only studies of individual treatment of
youth were included, 6 of the 16 studies provided a measure of the caregiver alliance. Overall,
findings so far support the assumption that both the caregiver and the child alliances are
related to the treatment process but that the way they contribute may be somewhat different
(Green, 2006; Hawley & Weisz, 2005; Zack et al., 2007). In particular, several studies have
found that the strength of the caregiver alliance is predictive of dropout and treatment
attendance (Garland, Haine-Schlagel, Accurso, & Baker-Ericzn, 2012; Hawley & Weisz,
2005; Shelef, Diamond, Diamond, & Liddle, 2005). However, no studies have investigated
how the caregiver alliance is related to the process and outcome in treatment of traumatized
youth, which is one of the aims of paper III.
13

1.4.5 Including the systemic context: interplay between youth and caregiver
perspectives. Some authors have argued that to fully understand how therapeutic
relationships influence the treatment process, the systemic context in which these alliances are
formed should be considered (Jensen et al., 2010; Pinsof & Catherall, 1986; Robbins et al.,
2006; Robbins et al., 2003), meaning that the interplay between the caregiver(s) and the
childs alliances should be investigated. Regardless if caregivers are directly involved in the
treatment or not, they are important agents for the childs meaning making and understanding
of the treatment process. This fact implies that caregivers will provide explicit or implicit
feedback on how the treatment should be understood and utilized (Jensen et al., 2010; Zack et
al., 2007). Most youth have little experience with therapy and may be unsure of what to
expect and how to feel about entering therapy. In a study of children exposed to sexual abuse,
it was found that the children made more or less explicit assessments of their parents
expressions of approval of the therapist and of the treatment, and this assessment influenced
the youths own thoughts about therapy. This process reflects what scholars call social
referencing, a strategy that children use to interpret their caregivers attitudes and emotions in
making meaning of a new situation (Campos, 1984; Feinman, 1992; Morris, Silk, Steinberg,
Myers, & Robinson, 2007). It can thus be expected that a weak parent-therapist alliance will
have a negative influence on the youths perception of the treatment and their own alliance to
the therapists. Conversely, one could also expect that if the child expresses strong negative
attitudes toward the therapist, these views will influence the caregivers view of their
therapeutic alliance.
This assumption resonates well with the Integrative Psychotherapy Alliance (IPA)
model (Pinsof & Catherall, 1986). This model was developed in order to transfer the alliance
construct into couples and family therapy, and there is an explicit focus on the mutual
influence between the alliances of the different family members in treatment. This interplay
between child and caregiver alliances may be particularly relevant to the treatment of
traumatized youth. Exposure to trauma may alter a childs core beliefs about the outside
world and other people, as well as influence their perceived ability to cope with future
challenges (Janoff-Bulman, 1992; Meiser-Stedman et al., 2009; Pynoos et al., 1999). This
perspective can make children more dependent on their caregivers reassurances. A child who
has been exposed to trauma within a family context may also be sensitive to his or her
parents feelings and whether he or she can discuss what has happened. Caregivers may also
struggle with their own emotions, such as shame, guilt, and distress (Davies, 1995; Deblinger,
Mannarino, Cohen, & Steer, 2006; Elliot & Carnes, 2001; Holt, Cohen, Mannarino, & Jensen,
14

2014; Kelley, 1990), and may therefore feel ambivalent about bringing their children to
therapy. How children interpret this ambivalence may in turn influence their own decision to
remain in treatment.
Studies based on the IPA model suggest that adults treatment attendance in individual
therapy was influenced by their perceptions of a non-attending partners acceptance of the
therapy (Pinsof, Zinbarg, & Knobloch-Fedders, 2008). Furthermore, at least one youth study
has found that the degree of parent-youth agreement about their relationship to the therapist
predicted dropout (Robbins et al., 2003). These findings point to the importance of learning
more about how the interplay between youth and caregivers is related to the therapy process.
In particular, youths perceptions of caregivers approval seem important. In order to
investigate this subject, a new scale was developed (paper III). The Child- and AdolescentPerceived Parental Approval of Treatment Scale (CAPPATS) included items that asked
whether the youth perceived that their parents liked the therapist and thought the treatment
was important and helpful, whether they thought that their parents wanted them to speak
openly about what they have experienced, and whether they thought their parents agreed with
them about the goals of the treatment (see the Appendix 1a and 1b).

1.5 Disentangling the Alliance-Outcome Relationship

Although the youth alliance field has made some progress during previous years and
the number of studies investigating the alliance-outcome relationship has grown, there is still
little knowledge of how the alliance may contribute to outcome (Crits-Cristoph, Gibbons, &
Hearon, 2006; Green, 2006; Karver, Handelsman, Fields, & Bickman, 2005). As pointed out
by Kazdin (2009), showing that alliance predicts later outcome by itself does not show that
alliance plays a causal or mediational role in therapeutic change. The concept of the alliance
is currently challenged by several methodological and conceptual problems that may question
the validity of the alliance as an important agent of therapeutic change (see e.g., DeRubeis,
Brotman, & Gibbons, 2005; Elvins & Green, 2008; Safran & Muran, 2006; Wampold & Imel,
2015). These are aspects that need to be addressed in order to better understand the allianceoutcome relationship and thus enable the research findings to help improve patient care
(Crits-Cristoph et al., 2006).
1.5.1 Challenges to the role of the alliance as an agent of therapeutic change. The
therapeutic alliance is often assumed to be either a mediator and/or a change mechanism in
therapy. A mediator is defined as a variable that may account for the relationship between an
IV and a DV, but it is not necessarily the cause of this relationship. In contrast, a change

15

mechanism is defined as the basis for the effect, i.e., the processes or events that are
responsible for the change (Kazdin, 2009). In both cases, the alliance can be assumed to be an
agent of therapeutic change.
One of the challenges to the role of the alliance as an agent of therapeutic change
regards client characteristics that may influence both the therapeutic alliance ratings and the
treatment process. As described earlier, it may be that the ability to form a relationship and
establish a good therapeutic alliance is influenced by the youths early attachment security
and prior relational experiences (Bowlby, 1988). Other relevant youth characteristics could be
initial symptom level, motivation for change, pretreatment expectancies, etc. (Elvins & Green,
2008). So far, these assumptions have been partially supported by studies reporting that client
variables such as pre-treatment social functioning, interpersonal style and symptom severity
influence the alliance formation in therapy with youth (Kazdin & Whitley, 2006; Levin,
Henderson, & Ehrenreich-May, 2012) and adults (see e.g., Crits-Cristoph et al., 2006;
Hersoug, Hglend, Havik, von der Lippe, & Monsen, 2009; Muran, Segal, Samstag, &
Crawford, 1994). Some of these client characteristics have also demonstrated a direct relation
to therapy outcome (Crits-Cristoph et al., 2006). However, a study analyzing therapist effects
on the alliance-outcome relationship found therapists contributions to the alliance to be a
stronger predictor of outcome compared to client characteristics (Del Re, Flckiger, Horvath,
Symonds, & Wampold, 2012).
Another aspect that may influence the alliance-outcome relationship is reverse
causation. Although it is commonly assumed that the alliance predicts symptom reduction, it
could also be that the alliance is the result of early treatment gains. To date, most studies of
the alliance in youth therapy have measured the alliance either late in treatment or at the same
time point as the outcome, making it difficult to rule out whether the alliance ratings are
influenced by early symptom relieves (Shirk et al., 2011). Although there is an increase in
studies reporting that the alliance predicts outcome even after controlling for early treatment
gains (Crits-Cristoph, Gibbons, Hamilton, Ring-Kurtz, & Gallop, 2011; Marker, Comer,
Abramova, & Kendall, 2013), establishing a causal relationship between alliance and outcome
in the future necessitates careful control of the contribution of symptom change to the alliance
(Elvins & Green, 2008).
In addition, there are methodological issues that can potentially threaten the validity of
the alliance-outcome association. One of these is related to the tendency to use the same rater
for both alliance and outcome, increasing the risk of a halo-effect, or shared method
variance. For example, in the McLeod (2011) analyses it was found that the alliance-outcome
16

relationship was significantly stronger when the same person was used to report both the
alliance and the outcome compared with studies where the alliance and outcome were
reported by different persons. In order to minimize this problem, studies should include
perspectives from different raters and assess the relationship between, for example, youthrated alliance and clinician-rated outcomes.
1.5.2. Potential pathways from the alliance to outcome. In spite of the
methodological and conceptual problems identified above, several authors argue that the
alliance is still an important concept worth retaining (Safran & Muran, 2006). However, in
order to strengthen this assumption, the field has to move forward to investigate the pathways
through which the alliance is responsible for change (Crits-Cristoph et al., 2006; Green, 2006;
Norcross & Lambert, 2011). To date, two theoretical frameworks of how the alliance is
related to outcome in youth therapy have been proposed. One is the model of the
mechanisms of the alliance outlined by Green (2006). This model is based on Hougaard
(1994) and his synthesis of the therapeutic alliance literature of adults dividing the alliance
into a personal alliance and a task-related alliance. According to Green (2006), the
mechanism of the personal alliance can be understood in light of Bowlbys (1988) attachment
theory. It is suggested that the establishment of a positive emotional bond will mobilize the
youths own coping resources, which will in turn alleviate illness-related anxieties and
distress. The second aspect of the model concerns the contractual nature of treatment and the
task-related alliance. Here, it is assumed that the process of negotiation and agreement on the
goals and tasks of the treatment will be beneficial in and of itself since it enables youth to feel
more empowered, motivated and involved in the treatment tasks. This model is useful since it
offers some plausible theories of how the youth alliance can contribute to change. However,
although specifically developed for youth therapy, it does not address to what degree the
caretaker or therapist alliances are related to the different pathways.
The second framework is the theoretical model of common process factors in youth
and family therapy developed by Karver and colleagues (2005).The aim of this model is to
link the various relationship variables present in youth treatment to outcome, and it is based
on an extensive review of theoretical and empirical studies. The model emphasizes how child,
caregiver and therapist characteristics and behaviors contribute to the child and familys affect
toward the therapist, their willingness to participate in the treatment, and their actual
involvement in the treatment tasks. It is then assumed that the bond, agreement and
involvement may contribute to the outcome of treatment in several different ways: 1) either as
a necessary relational change mechanism, 2) as a catalyst for other treatment processes that
17

lead to positive outcomes, or 3) as a moderator of therapist-offered interventions. This model

offers a more detailed framework of what influences the formations of the youth and
caregiver alliances, however there is little focus on the pathways from alliance to outcome.
There is, for instance, no mention of how the alliance can be a relational change mechanism
or in what way it can moderate the therapist-offered interventions.
A third relevant model has been developed in the adult field and this is the tripartite
model of relationships in psychotherapy by Wampold and Budge (2012). In this model, there
is first an initial relationship formation, followed by three relationship pathways. The initial
therapeutic bond refers to the early connection between therapist and client that has to be
established before the therapeutic work can be started. The subsequent relationship pathways
are assumed to involve different mechanisms of change, with the first referring to the real
relationship between therapist and client. This relationship is defined according to Gelsos
description of a relationship between a therapist and a client that is realistic (i.e., it is free
from transference) and genuine (i.e., it is authentic, open and honest) (Gelso, 2009). In their
model, Wampold and Budge show how this real relationship may offer belongingness and a
social connection that can help increase the patients quality of life. In the second pathway,
symptom reduction is achieved as a result of positive expectations that have been created
through explanation and some form of treatment. In particular, hopes and expectations that
the client will be able to cope with the difficulties that brought them to therapy are evoked,
and these expectations are assumed to be beneficial in and of themselves. It is of less
importance what kind of explanations and interventions are offered, as long as the client
accepts these explanations and participate in the therapeutic tasks assumed to be helpful. In
the third pathway, the relationship promotes the clients involvement in specific healthpromoting therapeutic tasks, which will again lead to symptom reduction. Here it is the tasks
performed (i.e., replacing maladaptive appraisals with more realistic and healthy ones,
engaging in social activities, or reducing perceived stress through relaxation exercises) that
are assumed to be health promoting, above and beyond the expectations created in the second
pathway. According to the authors it remains undetermined whether the benefit of these tasks
is because of specific factors, i.e., specific tasks or procedures that act as change mechanisms
in and of themselves for specific disorders, or whether there are rather a myriad of healthy
actions for numerous mental disorders.
In an attempt to better understand how the alliance may contribute to change in the
treatment of traumatized youths, a new model is proposed (Figure 1). This pathways from
the therapeutic alliance to outcome model integrates some of the features of the models
18

presented above, but has been adapted to fit in a youth treatment setting. Some of these
adaptations are including the caregiver alliance in the model, and assuming that the three rater
perspectives (youths, therapists and caregivers) may be differently related to the treatment
process and outcome. In line with the model of Karver et al., it is expected that youth,
caregiver and therapist pretreatment characteristics will influence the alliance formation. In
addition, ecological and contextual factors such as cultural believes, psychotherapeutic
treatment delivery systems and practical barriers are assumed to influence the treatment
process (see e.g., Orlinsky et al., 2004), but these factors not further elaborated in the model.
Based on the arguments by Shirk and Saiz (1992), the alliance is defined as an emotional
bond and agreement on treatment tasks. However, as a study by Accurso et al. (2013) indicate
that youths and caregivers do not divide the alliance into one task and one bond dimension,
but rather perceive the alliance more as a valence based construct with a positive and negative
dimension, the pathways model does not distinguish between how the emotional bond may
relate to outcome compared to the collaboration on the treatment tasks. The three suggested
pathways are assumed to be complementary and involve different, but not mutually exclusive,
mechanisms of change. Furthermore, the three pathways may be differently emphasized in
different treatment models, meaning that not all pathways need to be present in all types of
treatments.
Figure 1. Pathways from the therapeutic alliance to outcome

19

Similar to the model of Wampold and Budge (2012) it is assumed that an initial
therapeutic alliance is a prerequisite for the child and family to remain in treatment and be
involved in the therapeutic work. In this first step of the model, the alliance is therefore
expected to influence treatment attendance and dropout. The alliance may be particularly
important for treatment attendance in the treatment of traumatized youth, where attachments
may have been insecure and aspects of the therapy such as exposure may be perceived as very
challenging. One can thus expect the youth-therapist alliance to predict dropout. Furthermore,
the caregiver alliance may be important, since a strong relationship with the therapist can help
motivate caregivers who are potentially burdened by many life stressors to attend the
treatment sessions despite daily struggles and practical barriers. However, these barriers are
also assumed to have direct influence on treatment attendance, together with youth, caregiver,
and therapist pretreatment characteristics (as marked by the dotted lines in the model). The
importance of this first step has been partially supported by the studies discussed above that
find a link between caregiver alliance and dropout in treatment of other disorders (Garland et
al., 2012; Hawley & Weisz, 2005; Shelef et al., 2005), but less is known about the
relationship between the youth-therapist alliance and dropout.
Once the family is involved in the treatment, the therapeutic alliance may contribute
to change as a healing factor in and of itself. This is pathway 1 in the model and is similar to
what Green describes as the personal alliance and Wampold and Budge call the real
relationship. Although Gelso (2009) emphasize that the real relationship is not the same as a
working alliance, the two types of relationships are assumed to be highly overlapping and
may in practice be indistinguishable. Furthermore, since studies have shown that clients do to
a lesser degree differentiate the alliance from the real relationship compared with therapists
(ibid.), and youth and caregivers do not seem to differentiate the emotional bond from the
therapeutic tasks (Accurso et al., 2013), the therapeutic alliance is assumed to reflect the real
relationship in the current pathways model. The healing effect of the alliance as a real
relationship can be understood in light of Bowlbys attachment theory, where a personal
connection to another human being that is invested in ones well being is assumed to be health
promoting in and of itself (see Wampold & Budge, 2012). Furthermore, in a relationship
where the therapist is genuinely attuned to the client, the therapist may also be able to identify
and repair potential ruptures in the alliance that may occur during the treatment. The
reparation of such alliance ruptures is suggested to represent a corrective emotional
experience for the patient (Safran, Crocker, McMain, & Murray, 1990; Safran & Muran,
2000), and a link has been established between such corrective emotional experiences and the
20

ability to resolve early relationship traumas (Hartman & Zimberoff, 2004). More specifically,
it has been claimed that a therapeutic relationship can enhance the neural integration and
cortical circuitry of brain structures potentially damaged by early traumatization, which in
turn enhances emotional regulation and the ability to relate to a coherent trauma story (Siegel,
2003). In the pathways model, it is assumed that in particular the clients perspective on the
alliance may reflect the extent to which the youth or the caregiver feels connected to the
therapist. Furthermore, it can be expected that agreement between the client and therapist on
the alliance may reflect the level of attunement and the degree to which the two participants
have developed a real relationship.
The second pathway is similar to Greens task-related alliance and Wampold and
Budges creation of expectations pathway. In this pathway it is assumed that the alliance
contributes to change because the emotional bond and agreement on the task will facilitate the
clients involvement in the therapeutic procedures provided by the therapist. This involvement
will then increase the clients experience of being able to cope with their problematic thoughts
and feelings, creating hope and a sense of increased competency that contribute to change. In
this pathway, it is assumed that it is not important what type of task is performed as long as
the client and the therapist perceives the task as meaningful and relevant (Wampold & Budge,
2012). The relevance of this pathway has been supported by meta-analytic findings showing
that the alliance is a consistent predictor of outcome across a variety of treatment
interventions and tasks (Flckiger et al., 2012; McLeod, 2011; Shirk et al., 2011). In the
treatment of youths it may be important for the therapist to focus on the alliance both with the
youth and the caregiver(s) as the participants may hold different views of what types of tasks
may be helpful and contribute to an increased feeling of hope and competency.
The third pathway bears resemblance to pathway 2 where the alliance facilitates the
involvement in therapeutic tasks. However, in pathway 3 it is assumed that the tasks
performed are the actual contributors of change, above and beyond the hopes and expectations
created in the second pathway. Here it has to be considered that although some therapeutic
tasks can seem very different they may in fact involve the same underlying change
mechanisms. For example, both psychodynamic therapy and cognitive behavioral therapy will
emphasize the importance of confronting, instead of avoiding, anxiety laden material,
although they may have different strategies for this confrontation (Weinberger, 2014). This
does not mean, however, that any therapeutic tasks will be helpful (Asnaani & Foa, 2014). In
the treatment of PTSS one such specific task could be the creation of a trauma narrative. This
task includes several aspects that are assumed to reduce pathological processes that have been
21

linked to the development and maintenance of PTSS. For one, the creation of a coherent
narrative will contribute to an increased elaboration and contextualization of the traumatic
event in the clients autobiographical (see Ehlers & Clark, 2000). In addition, when the youth
is involved in talking and thinking about the traumatic event in great detail and over a period
of several sessions, this will represent a form for exposure that can help reduce the fearrelated associations connected to the traumatic memory. And last, working through a trauma
narrative can help identify and challenge potential negative attributions that the youth may
have related to the trauma. This may be beneficial as these appraisals have been found to
predict the development and maintenance of PTSS (Meiser-Steadman et al., 2009). Although
there is, as already mentioned, some controversy regarding the importance of specific factors
in treatment of adult PTSD (see Asnaani & Foa, 2014; Benish et al., 2008; Ehlers et al.,
2010), several studies have found that the inclusion of a trauma-specific component such as a
trauma narrative and/or exposure is associated with lower levels of post-treatment PTSS both
in the treatment of youth (Deblinger et al., 2011) and adults (Ehring et al., 2014; Powers,
Halpern, Ferenschak, Gillihan, & Foa, 2010). In this third pathway it can be assumed that the
youth alliance may be more important compared with the caregiver alliance, because the tasks
are specifically aimed at changing the youths thoughts and feelings.
In summary, it can be expected that both the first step and all of the three subsequent
pathways will be important in the treatment of traumatized youth, but that the relative strength
of the pathways may differ across treatment models (DeRubeis et al., 2005). For instance, in
more structured therapies where there is explicit emphasis on specific treatment tasks, it can
be assumed that the alliance is important as a facilitator for involvement in these activities. On
the other hand, in therapies where the focus is more on the therapeutic relationship, one can
expect that the alliance to be a healing factor in and of itself. In terms of the different
participants perspectives on the alliance, it can be assumed that youth, caregiver, and
therapist ratings of the alliance may predict outcome in the treatment of youth PTSS, but that
they may be related to different pathways. For instance, the caregiver alliance can be assumed
to be important for treatment attendance, but to a lesser degree related to the relationship
pathway (1) or the involvement in specific tasks pathway (3).
In this thesis, different aspects of the pathways model will be investigated in the three
papers. Collectively, the results may potentially contribute to an increased understanding of
the role the therapeutic alliance plays in the treatment of traumatized youths.

22

2. The Present Study

2.1. Aims and Research Questions
The primary aim of this thesis is to better understand the role the therapeutic alliance
plays in treatment of traumatized youth. More specifically, this work 1) investigates the
relationship between alliance and post-treatment symptom reduction in youth trauma
treatments; 2) elucidates the therapists perspectives of the alliance, and how these
perspectives were related to adolescents ratings, process and outcome; 3) looks at the
relational processes involved in premature treatment termination. More specifically, the
following research questions were pursued in this thesis:
1) What is the relationship between the therapeutic alliance and outcome in treatment of youth
suffering from PTSS? Is this relationship the same in TF-CBT and TAU? (paper I)
2) How are therapist ratings of the alliance related to process and outcome, and to what
degree do youth and therapist perspectives on the alliance overlap? Are there any differences
in the underlying factor structure of youth and therapist alliance ratings? Does the lack of
agreement across raters influence outcome? (paper II)
3) How are youth characteristics such as demographic variables and trauma history related to
dropout? Do treatment method and caregiver participation influence dropout? How do youth,
therapist and caregiver first session alliance ratings and youth-perceived parental approval of
treatment relate to dropout? (paper III)

3. Materials and Methods

3.1 Procedure: The Norwegian TF-CBT Study
The data in this thesis were derived from The Norwegian TF-CBT Study, a
randomized effectiveness study investigating effect and process variables in the treatment of
traumatized youth in community clinics. All participants were referred through standard
referral procedures (i.e., by the primary physician or child protective services) to one of the
eight participating child and adolescent mental health clinics. These clinics are situated in
different parts of Norway; four of the clinics are located in small cities, two are in large cities,
and two are in suburban areas. Youth between 10 and 18 years old, reporting exposure to at
least one traumatizing event at least four weeks prior to intake, and presenting PTSS above a
pre-established cutoff (RQWKHChild PTSD Symptom Scale (CPSS; Foa, Johnson, Feeny,
& Treadwell, 2001) were invited to participate. Exclusion criteria were acute suicidal
23

behavior, psychosis, intellectual disability or need for an interpreter. Recruitment took place
from April 2008 until February 2011 and a total of 454 youth were screened for eligibility. Of
these, 200 met the inclusion criteria, and 156 agreed to participate (Figure 2). Information
about the study was given both verbally and in written form, and written consent was obtained
from both the caretaker and the adolescent. Procedures were reviewed and approved by the
Regional Committee for Medical and Health Research (REC). After consent was given,
participants were randomized to receive either TF-CBT (n = 79) or the therapy normally
provided at the clinic (TAU, n = 77). Computer-generated randomized block procedures were
used, one for each clinic, and participants were not stratified on any specific features.
Symptom levels were assessed pre-treatment (T1a), after six sessions (T2), and posttreatment (T3). In addition, follow-up data were collected 1 year after the pre-treatment
assessment (T4) and 18 months after the post-session assessment (T5). In this thesis, data
from the three first time points are included. Since TF-CBT, normally delivered over a course
of 1215 sessions, was the experimental condition, it was decided that all post-treatment
assessments should be conducted after the 15th session, even though some participants in the
TAU and TF-CBT condition were still in treatment at this time point. Alliance ratings were
collected after sessions one (T1b) and six (T2), and youths perceptions of their parents
approval of the therapy was collected after session one (T1b). In order to reduce social
desirability, youth and caregivers were informed that their therapist would not be able to see
their ratings. Two licensed psychologists from the research group administered all of the
assessments, and they were blind to the youths treatment conditions. All assessments, except
the diagnostic interview, were conducted by computer-assisted self-report, but the
psychologists were available and could answer any questions if necessary. The caregiver
ratings were completed by the same caregiver each time. Participating adolescents received a
small gift card (e.g., a movie pass) after completing the post-treatment assessment (T3), but
no other economic compensation was provided.

24

Figure 2. Flowchart participants.

25

3.2 Sample
Table 1: Short overview of the main focus and different subsamples in each paper
Paper Main focus
Youth-rated alliance and relationship to
I
outcome across treatment conditions

Sample
156 youth

Analytical method
Hierarchical linear
regression analyses

156 youth
71 therapists

Linear regression
analyses, exploratory
factor analyses

II

Youth and therapist perspectives of the

alliance: relationships to outcome, process
and underlying dimension

III

Predictors of dropout. Youth, therapist and

156 youth
caregiver perspectives; background variables, 96 caregivers
treatment method, and first session process
71 therapists
variables

Binominal logistic
regressions

3.2.1. Youth sample. The total youth sample consisted of 156 adolescents (79.5%
girls), with a mean age of 15.1 years (SD: 2.20 years; range: 1018 years). The majority of
participants had at least one European-born parent (81.4%) and lived in one-parent
households (63.6%). Adolescents reported being exposed to an average of 3.6 different types
of traumatizing events (SD: 1.8; range: 110). The most frequently reported traumas were
violence or threats of violence outside the family context (75.0%), traumatic loss (i.e., sudden
death of a caregiver or a close person (60.9%), physical abuse by a caregiver (45.5%),
witnessing violence within the family (42.9%), and sexual abuse by someone outside the
family (30.8%). When asked to specify which event they perceived as the most disturbing or
severe (the worst event), at intake, 32.1% of youth reported exposure to domestic violence
and physical abuse, 28.8% reported sexual abuse, 17.9% reported traumatic loss, 17.3%
reported violent attacks outside the family context, and the remaining 4.0% reported accidents
or other forms of non-interpersonal traumas (Table 2). All participants reported clinically
elevated symptoms of PTS; mean-level CPSS scores at pre-treatment were 27.2 (SD: 7.7).
Based on clinical interviews with the Clinician-Administered PTSD Scale Child and
Adolescent (Nader et al., 2004), the majority of the sample (66.9%) satisfied the diagnostic
criteria for PTSD at intake. In addition, 72.8% of the youth scored above the clinical cutoff
for depression on the Mood and Feelings Questionnaire (Angold, Costello, Messer, &
Pickles, 1995), 66.4% scored above the cutoff for anxiety on the Screen for Anxiety-Related
Disorders (Birmaher et al., 1999), and 59.1% had other behavioral and attention problems, as
rated by the Strengths and Difficulties Questionnaire (SDQ) (Goodman, 2001).
26

Table 2: Participant characteristics

Youth-reported traumatic events, total exposure
Violence outside family
Traumatic loss
Physical abuse
Witnessing domestic violence
Sexual abuse outside the family
Witnessing violence outside the family
Severe accidents
Hospitalization
Sexual abuse inside the family
Other traumatic event
Source trauma
Physical abuse and domestic violence
Sexual abuse
Traumatic loss
Violence outside the family
Accidents/ hospitalization
War/ refuge
Ethnic background
Both parents Norwegian
One parent Norwegian
Asian
Other European
African
Latin American
Housing situation
With both parents
With one parent
Foster care
Alone/ other arrangements
Missing
Parent education
Elementary school
High school
Vocational training
College (14 years after high school)
University (> 5 years)
Missing

75.0
60.9
45.5
42.9
30.8
27.6
20.5
16.7
7.7
36.6

117
95
71
67
48
43
32
26
12
57

32.1
28.8
17.9
17.3
2.6
1.3

50
45
28
27
4
2

73.1
9.6
10.3
3.8
1.9
1.3

114
15
16
6
3
2

21.8
61.6
7.7
6.4
2.6

34
96
12
10
4

12.2
31.4
10.3
28.2
5.1
12.8

19
49
16
44
8
20

27

Total household income1

< USD $35,000
13.5
USD $35,00088,000
30.8
USD $88,000174,000
24.4
> USD $174,000
5.8
Did not know/ did not want to state
25.6
1
Mean income in Norway in 2012 = USD $79,800 (www.ssb.no)

21
48
38
9
40

3.2.2 Caregiver sample. A total of 135 caregivers participated in the main study. Of
these, 96 caregivers participated in the first session either alone or together with their child,
and these were included in paper III. This sample consisted of 64 biological mothers (66.7%),
17 biological fathers (17.7%), 12 foster parents (12.5%), and three (3.1%) caregivers that did
not report their relationship to the child. The majority of caregivers had no education after
high school (55.6%) and two-thirds of caregivers were working full- or part-time (64.4%).
Comparisons between the two samples that had participating and non-participating caregivers
showed that the youth in the latter group were significantly older (M age = 16.4 years vs. 14.2
years, t[1, 153.8] = 7.5, p < 0.001) and reported exposure to significantly more traumatic
events (M = 4.1 vs. 3.4, t[1, 102.5] = 2.4, p = .019), but there were no significant differences
in the pre-treatment CPSS levels or ethnic minority status of the groups.
3.2.3 Therapist sample. The participating youth were treated by 71 therapists. All
volunteered to participate in the trial and were not randomized to treatment conditions. In the
TF-CBT condition, 26 therapists were recruited and received training. Most therapists were
psychologists (80.8%, n = 21), 7.7% (n = 2) were psychiatrists, 7.7% (n = 2) were educational
therapists (Masters of education and additional clinical training), and 3.8% (n = 1) were
clinical social workers (a Bachelor-level degree with additional clinical training). On average,
therapists had 10.2 years of experience (SD: 6.4 years; range: 328 years), and in this study
the therapists treated on average 3.1 (Mdn: 3; SD: 1.5; range: 16) participants each. In regard
to the alliance ratings, 24 therapists reported their alliance after session six with an average of
2.6 youth (Mdn: 3; SD: 1.4; range: 15). When asked about theoretical orientation, 61.5% (n=
16) of the therapists characterized their background as cognitive-behavioral, 23.1% (n= 6)
characterized their background as psychodynamic and 7.7% (n= 2) characterized their
background as family/systemic. All therapists received between four and six days of initial
training, and were encouraged to read the treatment manual (Cohen et al., 2006) and complete
a web-based learning course for TF-CBT (www.musc.edu/tfcbt). Treatment adherence was
supported through initial session-by-session supervision provided by trained TF-CBT

28

therapists based on reviews of audio-recorded sessions. As the therapist became more familiar
with the model, supervision was reduced to bi-weekly sessions.
In the TAU condition, 45 therapists participated. The sample consisted of 51.1% (n =
23) psychologists, 26.7% (n = 12) clinical social workers, 17.8% (n= 8) educational
therapists, and 4.4% (n= 2) psychiatrists. Based on self-reporting, 37.8% (n= 17) of the
therapists described their theoretical orientation as psychodynamic, 24.4% (n= 11) described
their theoretical orientation as cognitive-behavioral and 20.0% (n= 9) described their
theoretical orientation as family or systemic. The remaining eight therapists did not report
their theoretical orientation. The mean work experience in this group was 12.5 years (SD:
10.3 years; range: 140 years), and they treated on average 1.7 (Mdn: 1; SD: 1.2; range: 18)
participants each. Thirty-eight therapists reported their alliance after session six with an
average of 1.7 youth (Mdn: 1; SD: 1.1; range: 16). Therapists in the TAU condition reported
receiving on average 1.4 hours of supervision (SD: 5.3; range: 040) on their therapies with
study participants in total.

3.3 Treatment Conditions

3.3.1 TF-CBT. As described in the Introduction (page 8), TF-CBT is a short-term,

component-based, manual-guided treatment developed by Cohen, Mannarino, & Deblinger

(2006). In this study, participants in the TF-CBT conditions had completed on average 13.0
(SD: 3.1; Mdn: 14; range: 417) sessions before the T3 assessment, and on average 18.8 (SD:
8.4; Mdn: 16.0; range: 849) sessions before the case was discharged from the clinic. The
therapy sessions were audio-recorded, and each session was coded for fidelity by trained TFCBT therapists, using a treatment adherence checklist provided by the treatment developers.
In cases where there were questions about fidelity, these were discussed, and fidelity was
determined by consensus. In five cases, the core TF-CBT components (i.e., psycho-education,
relaxation, emotion regulation, trauma narrative and cognitive restructuring) were not
provided. Since an explicit aim of paper I was to investigate the interaction between treatment
condition and alliance, these cases were excluded from the analyses.
3.3.2 TAU. In the TAU condition, therapists were not given any specific instructions
but were asked to provide the treatment they believed to be effective for each particular case.
Treatment was for the most part provided individually, but in 55.3% (n = 42) of the cases
parents were also involved in the therapy process. All TAU sessions were recorded, and at
least five sessions for each case were checked (1st, 2nd, 3rd, 6th and 9th, mean: 5.7 sessions; SD:
3.5; range: 119) with the TF-CBT fidelity checklist. In those cases where treatments had

29

features resembling TF-CBT, additional sessions were checked, adding up to a total of 392
fidelity-checked sessions. In addition, two or three sessions were drawn randomly from each
therapist (n = 81 sessions) and coded according to the Therapy Process Observational Coding
System Strategies Scale (TPOCS-S; McLeod & Weisz, 2010). Based on the fidelity
checklist, the following TF-CBT components were provided in the TAU condition: 15.6% (n
= 12) psycho-education, 11.6% (n = 8) affective expression and modulation, 8.7% (n = 6)
relaxation skills and 7.2% (n = 5) cognitive restructuring. In 5.8% (n = 4) of the cases, there
was some therapeutic work addressing the traumatic event but none of the cases included the
parents in the trauma work. None of the cases satisfied the adherence criteria for TF-CBT.
Codings with the TPOCS-S showed that the main strategies used in the TAU condition were
client-centered (present in 92.6% of the sessions) and psychodynamic strategies (present in
47.5% of the sessions). Family therapeutic interventions were observed in 35.8% of the
sessions, cognitive strategies were observed in 30.9% of the sessions, and behavioral
strategies were observed in 19.8% of the sessions. On average, participants in the TAU
condition had 12.9 (SD: 4.8; Mdn: 15; range: 121) sessions before the T3 assessment, and an
average of 23.8 (SD: 21.4; Mdn: 19; range: 1114) sessions before the case was discharged
from the clinic.

3.4 Measures
3.4.1 Youth alliance. The TASC-r (Shirk & Saiz 1992; Shirk & Karver, 2010)
consists of 12 items that measure both the emotional aspects (bond, items 1, 3, 5, 6, 8 and 10)
and degree of youth-therapist collaboration (task, items 2, 4, 7, 9, 11 and 12). The items are
worded as statements regarding the youths feelings toward the therapist (e.g., I like my
therapist) and their self-perceived involvement in tasks (e.g., I work with my therapist to
solve problems in my life), and all items are answered on a 4-point Likert scale (Not at all to
Very much). The TASC-R was translated and back translated, and the first author of the
TASC-R approved the Norwegian version. The internal reliability of the scale was good in
this sample both at T1 (WRWDOVFDOH bond = .82, task = .73) and T2 (total scale  
.91, bond = .88, task = .81).
3.4.2 Caregiver alliance. A parallel caregiver-therapist alliance form was developed
in collaboration with the first author of the TASC-r. This scale was to a large degree similar to
the caregiver version developed by Hawley & Weisz (2005). Items correspond to the child
scale, but my therapist was changed to the therapist I see at the clinic. This change
ensured that the form could include both those parents that met with their childs therapist,

30

and those who had parallel sessions with another therapist. In line with the youth and therapist
versions, all items were answered on a 4-point scale. In this sample, the internal consistency
RIWKHWRWDOVFDOH .74) and the bRQGVXEVFDOH were adequate, but the task
VXEVFDOH KDGDQDOSKDYDOXHEHORZWKHUHFRPPHQGHGOHYHO(Tavakol & Dennick,
2011).
3.4.3 Therapist alliance. The therapist versions of the TASC-R consist of the same 12
items as the youth and caregivers scales, only phrased so that the therapists rate their
impression of the youths or caregivers engagement (e.g., The child expresses positive
emotions toward you, the therapist; The child finds it hard to work with you on solving
problems in his/her life; The childs caregiver works with me to solve problems in his/her
life). It thus involves ratings of the clients bond and task involvement rather than the
therapists own. In this sample, the two scales internal consistencies were good (T1 therapist\RXWKWRWDOVFDOH 88, bRQG .86, task = .80; T2 therapist-\RXWKWRWDOVFDOH .91, bond
 tDVN DQG7therapist-FDUHJLYHUWRWDOVFDOH .83, bond = .75, task = .75).
3.4.4 Youths perceptions of parental approval of treatment. Based on earlier
studies (Brookman-Frazee, Haine, Gabayan, & Garland, 2008; Jensen et al., 2010; Pinsof et
al., 2008), a new scale was developed as part of this thesis. The aim of this scale, the Childand Adolescent-Perceived Parental Approval of Treatment Scale (CAPPATS), was to
investigate whether the youth participants believed that their parents approved of the
treatment. The CAPPATS consisted of five items that were worded so that the child could rate
his or her impression of his or her parents approval (1: I think my caregivers (e.g., mother/
father/ foster parent) like my therapist, 2: I think my caregivers think it is important that I
attend the sessions at the clinic, 3: I think my caregivers want me to speak openly about my
experiences to my therapist, 4: I think my caregivers and I agree on what problems to work
on, and 5: I think my caregivers think that the things the therapist and I do are helpful). All
items were rated on a 4-point scale (Not at all to Very much). To investigate the performance
of the new scale CAPPATS, we performed several analyses. First, the internal consistency (
= .74) was adequate. Furthermore, analyses of the response distributions of the individual
items showed that there was adequate variance in this sample (mean scores: 2.903.67;
skweness: -1.930.57), and that the average interitem correlation (.38) was within the
recommended level (Clark & Watson, 1995). Last, principal-axis analysis showed that all
items had loadings of .55 or higher, and that the one five-item factor explained 50.1% of the
variance. Taken together, these results showed that the CAPPATS scale performed well
according to the recommended criteria.
31

3.4.5 Trauma exposure. To assess trauma exposure, the research group developed a
checklist based on the items described in The Traumatic Events Screening Inventory for
Children (TESI-C; Ribbe, 1996), which included the following experiences: 1) severe
accident; 2) natural disaster; 3) sudden death or severe illness of a close person; 4) extremely
painful or frightening medical procedures; 5) violence or threats of violence outside the
family context; 6) robbery or assault; 7) kidnapping; 8) witnessing violence outside the
family; 9) witnessing violence within the family; 10) physical abuse within the family; 11)
sexual abuse outside the family; 12) sexual abuse within the family; and 13) other frightening
or overwhelming experiences. The checklist was administered as an interview by clinically
trained therapists, and a traumatic event was rated as present if the child reported that he or
she had felt scared, terrified, or helpless during or immediately after the event.
3.4.6 Self-reported PTSS. Adolescents PTSS were first measured by means of
the self-completion Child PTSD Symptom Scale (CPSS)(CPSS; Foa et al., 2001). The CPSS
consists of two parts. The first part measures the 17 symptoms of PTSD defined in the DSMIV, covering the three factors Re-experiencing, Avoidance and Hyper arousal. Symptom
frequency is rated based on the last two weeks, with a 4-point scale ranging from Never or
once to Almost every day. The second part measures how the symptoms impact daily
functioning, covering friendships, family, school work, hobbies and activities, house chores
and general life satisfaction. Principal component analyses of a comparable sample of 312
youth confirm the factor structure in the original version (Hukkelberg & Jensen, 2011), and
satisfactory internal consistencies were found for each of the three factors (Re-H[SHULHQFH 
$YRLGDQFH +\SHUDURXVDO DQGWKHfunctional impairment scale  
The scale was translated and back translated, and the developers of the scale approved the
Norwegian version.
3.4.7 Clinician-rated PTSS. In addition to the self-report measure, a clinicianadministered PTSD interview was conducted (CAPS-CA; Nader et al., 2004). The CAPS-CA
is a structured interview that assesses the frequency and intensity of the 17 DSM-IV-defined
symptoms of PTSD, and it is adapted from the adult version to be suitable both for younger
children and adolescents up to age 18. Items are scored on 5-point frequency scales (i.e., from
0 = None of the time to 4 = Most of the time) and 5-point intensity rating scales (i.e., from 0 =
Not a problem to 4 = A big problem, I have to stop what I am doing), assessing the symptom
frequency and intensity during the past month. Items are scored based on both the youths
answers and clinical judgment. The interview was translated and back translated, and the first
author of the CAPS-CA approved the Norwegian version. The entire scale showed
32

VDWLVIDFWRU\LQWHUQDOFRQVLVWHQF\ DVGLGWKH'60-IV-defined tripartite model (ReH[SHULHQFLQJ .87, AYRLGDQFH .77, H\SHUDURXVDO ,QWHU-rater reliability for
total sum score was excellent (ICC = .99; 95% CI: .951.00) and the kappa value of the
diagnostic status was .80.
3.4.8 Caregiver-rated PTSS. Caregivers perspective of the youths PTSS was
assessed using The UCLA PTSD Index for DSM-IV Parent Report Version (The UCLA Index;
Steinberg, Brymer, Decker, & Pynoos, 2004). The symptom checklist includes 21 items, 17 of
which are defined by the DSM-IV and are used to compute the sum score. Items 14 and 20
are not included in the scoring, and only the highest scores on either items 3 or 20 and the
highest scores of items 10 or 11 are used. Caregivers are asked to report the frequency of their
childs symptoms during the last month. Items are rated on a scale ranging from 0 (None of
the time) to 4 (Most of the time). In addition, a fifth option is given (5 = Dont know). In this
study, all ratings of 5 were coded as missing, and total scores were calculated for those
caregivers who had reported 04 on at least 9 of the 17 items (mean scores 17). The internal
consistency for the scale was good in this sample ( 
3.4.9 Depressive symptoms. The Mood and Feelings Questionnaire (MFQ; Angold et
al., 1995) was used to assess depressive symptoms. This self-report questionnaire was
designed to assess depressive symptoms in children and adolescents between 8 and 18 years
old. The questionnaire consists of 34 questions measuring both the full range of DSM-IV
diagnostic criteria for depressive disorders, as well as additional items reflecting common
affective, cognitive and somatic features of childhood depression. Items are scored on a 3point scale ranging from 0 (Not true) to 2 (True). In this sample, the MFQ showed good
LQWHUQDOFRQVLVWHQF\ 7KHVFDOHKDVEHHQWUDQVODWHGDQGEDFNWUDQVODWHGDQGWKH
Norwegian version was approved by the developers (Sund, Larsson, & Wichstrm, 2001).
3.4.10 Anxiety symptoms. The Screen for Child Anxiety-Related Disorders
(SCARED) is a self-report questionnaire developed by Birmaher et al. (1999). It measures
anxiety symptoms in children and adolescents aged 818 years. The instrument consists of 41
items that cover five specific anxiety disorders: 1) Panic Disorder or Significant Somatic
Symptoms; 2) Generalized Anxiety Disorder; 3) Separation Anxiety Disorder; 4) Social
Anxiety Disorder; and 5) Significant School Avoidance. Items are scored on a 3-point scale
ranging from 0 (Not true) to 2 (True). In this sample, the SCARED showed satisfactory
LQWHUQDOFRQVLVWHQF\RQWKHWRWDOVFDOH 7KHVFDOHZDVWUDQVODWHGDQGEDFNWUDQVODWHG
and the developers approved the Norwegian version.

33

3.4.11 General mental health. The Strengths and Difficulties Questionnaire

(SDQ)(SDQ; Goodman, 2001) is a self-report questionnaire measuring general mental health
problems in children and adolescents. The SDQ contains 25 items, covering five areas of
clinical interest: hyperactivity/inattention, emotional symptoms, conduct problems, peer
relation problems and pro-social behavior. Symptoms are reported based on the last six
months and the items are scored on a 3-point scale ranging from 0 (Not true) to 2 (True) for
the positively worded items, and 0 (True) to 2 (Not true) for the negatively worded items. The
total score of general difficulties is based on the four problem-oriented sub scores. The
authorized translated version of the SDQ was used (www.sdqinfo.com) and the scale showed
VDWLVIDFWRU\LQWHUQDOFRQVLVWHQFLHVRI 
3.4.12 Youth-rated treatment satisfaction. To rate youth satisfaction with the
therapy, a three-item self-report measure was developed. Items included I liked going to the
clinic, Going to the clinic helped me with my problems, and If I were ever having
problems again, I would want to come back to this clinic. All items were rated on a 4-point
scale form ranging from 1 (Not at all) to 4 (All of the time), and the scale was administered at
the post-treatment assessment. The internal consistency of the scale was good ( 

3.5 Statistical Analyses

3.5.1 Initial analyses (papers I, II & III). Comparisons between groups (i.e.,
differences between the non-randomized therapists in the two conditions, youth with and
without data from specific assessment points, cases with and without caregivers involved,
differences between youth and therapist ratings of the alliance, etc.) were computed using
independent sample t-tests, chi squared tests, analyses of variance (ANOVA), and paired
samples t-tests. Relationships between relevant variables were investigated with Pearson
correlations (r). Following Cohen (1992), r is a small effect when at least .10; a medium
effect when at least .30; and a large effect when at least .50. To test differences in
relationships between two conditions, bootstrap BC a intervals were used with 10,000
bootstrap replications. In addition, independent samples effect sizes (ES) were calculated
(Cohenss d: mean difference/ pooled SD).
3.5.2 Hierarchical regression analyses (papers I & II). Regression analyses are a set
of statistical analyses that allow one to assess the relationship between one dependent
variable (DV) and one or more independent variables (IVs). These analyses were used to
investigate the relationship between the therapeutic alliance (IV) and outcome (DV) in papers
I and II. Because of the nested nature of the data (youth nested within therapists and clinics),

34

using multi-level analyses is recommended to account for the non-independence in the data
(see e.g., West, 2009). To evaluate level of dependency in our data, the within-subjects intraclass correlations (ICC) were investigated. The results showed that the variance by clinic was
ignorable, with ICC variables ranging from 0.010.05 (Dyer, Hanges, & Hall, 2005). The
between-therapists ICCs ranged from 0.100.24, with an average ICC of 0.17. These values
were in the borderline range, implying that a two-level model including the therapist level
should be tested. However, with ICC levels in this low range there may be problems with
model stability (ibid.). Analyses were first run with linear mixed effects models (LME) with
youths nested within therapists. This resulted in reasonable estimates and mostly
interpretable loadings, but the confidence intervals were extremely large. The size of the
confidence intervals showed that models were essentially too unstable, and as a consequence
single-level regression analyses were used.
In paper I, another aim involved investigating any potential interaction effects between
treatment condition and the alliance-outcome relationship, and the predictors were entered
hierarchically in two steps. In the first step, pre-treatment scores, group condition and
alliance scores were entered, and in the second step an interaction between alliance scores
and group condition was included. For each step, the R2 was calculated, indicating the
amount of variance explained by the predictors in the model. This technique means that one
can calculate the additional information that is gained by adding new IVs to the model (Field,
2009).
3.5.3 Exploratory Factor Analyses (paper II). The factor structure of the youth and
therapist ratings of the TASC-r was investigated using EFA. Although an established
theoretical model of how the dimensions of the alliance would cluster (i.e., into a bond and
task dimension) exists, the aim of this study was to recover an empirical description of the
relationship between the items. EFA is the best method for this purpose (Tabachnick & Fidell,
2007). Geomin factor loadings were used with oblique rotation (Muthn & Muthn, 19982012). Since it was expected that two dimensions would emerge extractions were specified to
one or two factors. In order to determine the best model fit, both likelihood-ratio chi squared
(F2) and descriptive fit indices were utilized. The descriptive fit indices included the
comparative fit index (CFI), the root-mean-square error of approximation (RMSEA) and the
standardized root-mean-square residual (SRMR). These fit indices have most frequently been
studied as indicators of structural equation modeling and confirmatory factor analyses (see
e.g., Schreiber, Stage, King, Nora, & Barlow, 2006) and there are currently no established
cut-off values for the use of the fit indices in EFA (Barendse, Oort, & Timmerman, 2015). In
35

our study, we chose to follow Accurso et al. (2013), where models that fit very well (or
adequately) were LQGLFDWHGE\&),V0.95 (0.900.94), RMSEAs < 0.05 (to 0.08) and SRMR
< 0.05 (to 0.08). A model was assumed to be well fit if two of the three descriptive indices
indicated a good fit.
Multi-level analyses were also conducted in an attempt to account for the nested
nature of the data. However, similar to the LMM analyses the use of multi-level EFA resulted
in unstable models, and all analyses were run using single-level EFA.
3.5.4 Logistic regressions (paper III). To investigate the predictors of dropout in
paper III, logistic regression analyses were performed. These analyses allow for the
prediction of a dichotomous outcome such as group membership (e.g., dropout yes/no) when
predictors are continuous, discrete, or a combination of the two (Tabachnick & Fidell, 2007).
Regression models make no assumptions about the distribution of predictor variables,
meaning that they do not have to be normally distributed or of equal variance within each
group. The estimation results are expressed in terms of odds ratios (OR), which indicate the
change in odds resulting from a unit change in the predictor, or the change in group
membership (Field, 2009). ORs greater than 1 reflect an increase in odds of e.g., a certain
group membership; ORs less than 1 reflect a decrease in odds. Due to data being missing, the
logistic regressions were run with and without multiply imputed data (see Section 3.5.5).
3.5.5 Handling missing data (papers I, II and III). In all three papers, there were
missing data on several of the IVs. Missing rates were not significantly different in the two
treatment conditions, but the non-completers differed from the completing participants in
terms of several demographic and process variables (age, number of traumatic events and
alliance scores mid-treatment). This indicated that the data could not be assumed to be
missing completely at random (MCAR), and discarding data using list-wise deletion could
have increased the risk of obtaining a biased result (Schafer & Graham, 2002). In addition, the
high proportion of missing data pattern represented a substantial loss of power, since some of
the models had missing on more than 2/3 of the IVs. In order to investigate the potential bias
caused by the non-random missing data and reduced power, several steps were taken. First, in
all papers, the analyses were repeated using multiple imputation (200 completed data sets).
These analyses were then compared with analyses run with only the complete cases. In
addition, in paper I, the analyses were also run with weighted regression analyses. The
weighting model was based on a logistic regression for valid endpoint, with age, total number
of traumas reported, and either session one or mid-treatment alliance scores as covariates,
respectively. The results showed that the outcomes were comparable to the complete-case
36

analyses, strengthening the assumption that the missing data did not substantially bias our
results.
3.5.6 Statistical software. All initial analyses were conducted with PASW Statistics
19.0 (IBM SPSS Statistics, 2011), and CFA analyses were ran in Mplus 7.0 (Muthn &
Muthn, 19982012). The hierarchical regression analyses and logistic regressions with
imputed data were performed in R (The R Foundation for Statistical Computation, Vienna,
Austria), as well as the LME analyses. The multiple imputations were calculated using the R
package mice,bootstrapping with the R package boot, and LME with the R package
nlme.

3.6 Ethical Considerations

Sometimes there can be a conflict between what a researcher thinks is beneficial in
terms of knowledge acquisition and what can feel upsetting or offensive for a single
participant. Therefore, it is always important to consider whether participation in research can
be potentially harmful to a subject. In research projects that include children and adolescents,
this concern is particularly crucial, since these individuals may be less able to address their
needs compared with adults. In this study, all of the children and adolescents had been
exposed to at least one traumatizing event and presented high levels of psychological distress
upon intake. It was therefore important to carefully consider whether the assessments of
potentially traumatizing events and their consequences would represent an additional burden
on the affected youth. It has commonly been assumed that youth find it distressing to talk
about traumatic events, and historically research projects including traumatized youth have
had difficulties obtaining approval from institutional review boards because the work has
been considered to be too demanding (Dyb, 2007). However, studies so far have shown that,
when adequately organized, both traumatized youth and their caregivers report that they do
not feel overwhelmed or negatively affected by participation in trauma research, and that they
on the contrary find the interviews to be both useful and interesting (Griffin, Resick, Waldrop,
& Mechanic, 2003; Kassam-Adams & Newman, 2005).
To minimize the stress on the participants in this study, several precautions were
taken. First, in order to ensure that participation was voluntary and that the youth were aware
that they could withdraw from the interview or research project at any time point, information
was given both written and orally, and informed consent was obtained from both caretakers
and the youth. It was furthermore stressed that if any of the participants chose to withdraw
from the research project, doing so would not influence the quality or amount of therapy

37

provided. Secondly, clinical psychologists who were experienced in talking to children and
adolescents in difficult life situations conducted all of the interviews and assessments. Third,
all of the collected data were handled carefully so that the youth could feel safe that sensitive
information about them would remain confidential, unless issues of considerable concern (i.e.,
severe suicide ideation) were raised during the interview. Furthermore, the youth were
allowed to take breaks if needed and the assessment setting was adjusted in order to meet
individual needs, if required. Lastly, we also included questions about the assessment
procedures in the post-treatment measurement. The reason for this process was two-fold.
First, these questions provided the participants with an opportunity to give us direct feedback
about the assessment procedures to help us address potential weaknesses during the datacollection period. Secondly, reports from participants provided us with important information
about the potential strains experienced by youth and caregivers when they participated in the
research, which will help us guide future studies.

4. Results
4.1 Paper I: The Therapeutic Alliance in Treatment of Traumatized Youth:
Relation to Outcome in a Randomized Clinical Trial
In the first paper, the relationship between the youth-rated therapeutic alliance and
outcome was investigated. The alliance was assessed after sessions one (T1b) and six (T2)
and its predictive value was investigated in relation to a variety of outcome measures. These
relationships were evaluated both in the total sample and across the two treatment conditions
(TF-CBT and TAU).
The results showed that the alliance scores were comparable across treatment
conditions. However, youth receiving TF-CBT reported significantly fewer symptoms
compared with TAU after 15 sessions (T3). The analyses furthermore showed that the alliance
assessed at T2 was a significant predictor across different symptoms measures (PTS selfreported symptoms, PTSD diagnostic interview, depression, anxiety and general mental
health) in the TF-CBT condition, but these relationships were not present in TAU.
Furthermore, it was found that the T2 alliance ratings were not influenced by early symptom
change, but rather that the alliance at mid-treatment predicted subsequent change in
symptoms at T3.

38

This study is the first to investigate the contribution of alliance to outcome among
adolescents with post-traumatic symptoms, treated with TF-CBT or TAU. Our findings
indicate that there is an important interaction between alliance and therapeutic approach;
alliance predicted outcome in TF-CBT, but not in the non-specific treatment condition. A
positive working relationship appeared to be particularly important in the context of this
evidence-based treatment, which requires youth involvement in specific therapy tasks.
Furthermore, the findings showed that the use of a manual did not compromise alliance
formation.

4.2 Paper II: Therapist and Client Perspectives on the Alliance in the Treatment
of Traumatized Adolescents
In the second paper, adolescents and therapists ratings of the alliance were
investigated as predictors of outcome were investigated. In addition, the association between
client and therapist perspectives was explored, as well as the underlying dimensions of these
two perspectives. Finally, the level of discrepancy in youth and therapist ratings was
investigated as a predictor of outcome.
The results showed that both adolescent and therapist ratings of the alliance predicted
adolescents treatment satisfaction, but only the adolescent perspective was significantly
related to post-treatment symptoms. The level of adolescent-therapist agreement on the
alliance was moderate (ICC = .54, p < .001), with youth reporting on average a higher
alliance compared with therapists (t[99] = 6.5, p < .001). Factor analyses revealed differences
in factor structure with therapist ratings organized around bond and task dimensions and
youth ratings organized by item valence (i.e., whether the items were positively or negatively
worded). The discrepancy in youth and therapist ratings were significantly related to outcome.
Higher therapist ratings compared with youth ratings predicted higher residual PTSS and
lower treatment satisfaction.
Although adolescent and therapist alliance ratings are moderately associated, the
results suggest that the ratings are differentially associated with outcomes. These findings,
along with results indicating important differences in factor structure, imply that youth and
therapist ratings are not interchangeable. Future studies should investigate how therapists can
improve their judgments of youths perceptions of the alliance, since an overestimation of the
quality of the relationship seems to be negatively related to outcome.

39

4.3 Paper III: Understanding Dropout in the Treatment of Traumatized Youths:

Background, Treatment, and First Session Process Variables
Dropout is a common problem in community clinics and in the treatment of
traumatized youth in particular. In paper III, the aim was to better understand who is most
vulnerable to dropping out and why these clients quit. The work focused on examining
background variables, treatment method and first session process variables as potential
predictors of dropout. Perspectives from youth, caregivers and their therapists were
investigated, and dropout was defined based on therapist judgments.
The results showed that 43 (27.6%) participants dropped out, and that there were no
differences in dropout rates across the treatment conditions. Both youth and caregivers
reported alliance scores at the high end of the scale. Dropout was predicted by a lack of
caregiver participation, lower rates of youth-perceived parental treatment approval, and
weaker therapist-rated alliance ratings, but not by treatment condition, trauma exposure or
pre-treatment youth characteristics. Furthermore, neither youth nor parent ratings of the
alliance predicted dropout. Youth-perceived parental approval remained a significant
predictor, even after controlling for youth age, but this variable was not related to caregivers
own reports of their alliance to the therapist.
The findings from this study indicate that relationship variables are more important
predictors of dropout than treatment mode and youth background characteristics.
Furthermore, youth seem to place more weight on their caregivers approval of the treatment
compared with their own initial alliance to the therapist. This finding opens up pathways for a
new perspective on how the therapeutic alliance should be conceptualized. The results also
shed light on relational aspects beyond the youth-therapist dyad that may influence the
treatment process. The result that there were no differences in dropout rates across treatment
conditions is consistent with findings from adult studies, indicating that youth are able to
tolerate a trauma-focused and exposure-based treatment.

5. Discussion
The number of studies investigating the effect of therapeutic interventions for
traumatized youth is growing, providing important knowledge about treatments that can help
alleviate youths posttraumatic symptoms and distress. There are, however, still many
unanswered questions regarding how these interventions work and the processes contributing
to change. Although a strong therapeutic alliance is often described as an important success

40

factor in youth trauma treatment (Eltz et al., 1995; Lawson, 2009; Shirk & Eltz, 1998),
alliance has so far received only minimal empirical attention. Learning more about what role
the alliance plays in the process and outcome of therapy is an important step along the way to
further improve the treatment provided to trauma-affected youth and their families. In this
thesis, the therapeutic alliance was studied in the context of a randomized controlled trial, a
design that made it possible to explore several aspects of the alliance such as its relationship
to outcome and dropout, and make comparisons across two different treatment conditions. In
addition, the role of the alliance was investigated across different perspectives since therapist,
youth and caregiver ratings were included. The following discussion focuses on how the
findings from the three studies jointly contribute to increasing our understanding of the role of
the therapeutic alliance in youth trauma treatments.

5.1 Discussion of Main Findings

5.1.1 The alliance is a significant predictor of treatment process and outcome.
Taken together, our findings indicate that the therapeutic alliance is a central construct in the
treatment of traumatized youth, since it is significantly related to both outcome and treatment
process. The results from papers I and II show that a strong alliance was significantly related
to lower symptom levels post-treatment. This finding was in line with expectations and
corresponds with an early treatment study of maltreated youth (Eltz et al., 1995) and studies
of traumatized adults (Cloitre, Koenen, et al., 2002; Cloitre et al., 2004; McLaughlin et al.,
2013). Second, a strong alliance was significantly related to higher levels of treatment
satisfaction (paper II), a result that is similar to findings of studies of other clinical youth
populations (Hawley & Weisz, 2005; Kazdin et al., 2005). Although some authors have
argued that treatment satisfaction is not a good indicator of clinical improvement (Garland,
Aarons, Hawley, & Hough, 2003), it may still represent a relevant aspect of youths
experience with the therapy. Third, our findings suggest that a strong initial alliance may
reduce the risk of dropout (paper III), a finding that is consistent with the results of other
clinical child (Garland et al., 2012) and adult populations (Roos & Werbart, 2013). Lastly, we
found that the level of discrepancy in youth and therapist ratings predicted outcome (paper II).
Specifically, the findings revealed that when the therapist rated the alliance higher compared
with the youth, poorer outcomes ensued. This result resonates well with clinical literature
(Safran & Muran, 2000) and at least one study (McLaughlin et al., 2013) that emphasize the
importance of therapists ability to detect and repair ruptures in the alliance.

41

However, the results also indicate that several aspects influence the role of the
alliance. For one, it seems like treatment method moderate the relationship between alliance
and outcome, since the alliance predicted outcomes only in TF-CBT and not in TAU. Since
the results showed there were no differences in variance or overall level of alliance scores
between the two conditions, this finding indicates that there is an interaction between the
alliance and treatment method that is responsible for change. Such a synergistic effect has
been described in the adult literature (DeRubeis, 2005; Norcross & Lambert, 2011), and
differences across treatment conditions have been reported in other youth clinical populations
(Cummings et al., 2013; Hogue et al., 2006). However, this study was the first to investigate
this relationship in a youth trauma setting. Second, the findings suggest that different rater
perspectives predict different aspects of the outcome and process of treatment. In this thesis, it
was only the youth-rated alliance that predicted outcome; therapist ratings were not
significant. Conversely, therapist-rated alliance predicted dropout, but youth or caregiver
ratings were not significant predictors of dropout. This corresponds to the meta-analysis by
McLeod where rater perspective significantly moderated the alliance-outcome relationship
(McLeod, 2011). Given that the treatment participants differ in their degree of maturation,
power and treatment experience, it seems reasonable that they also play different roles in the
treatment process. The results suggest that youth, therapist and caregiver ratings of the
alliances are not interchangeable, but that they may instead provide information that is related
to different aspects of the treatment process.
Some of our expectations were not confirmed, however. In particular, given the
important role that caregivers play in youth treatments (Kazdin, Holland, & Crowley, 1997;
Nock & Ferriter, 2005), it was expected that the caregiver alliance would be a significant
predictor of dropout. However, this expectation was not confirmed (paper III). Since one
cannot draw firm conclusions from non-significant results, potentially confounding factors
should be evaluated. Other studies that showed that a weak caretaker-therapist relationship
was related to dropout assessed the alliance either late in treatment (Accurso et al., 2013;
Garland et al., 2012) or post-treatment (Garcia & Weisz, 2002; Hawley & Weisz, 2005;
Kazdin et al., 1997). It is possible that the first session was too early for the caregiver to
reliably report their alliance with the therapist. Alternatively, it might not be the first
impression but rather the change in alliance over time that is related to dropout. Therefore,
based on this thesis, it is still an open question whether and how the caregiver alliance is
related to the treatment process of youth experiencing trauma treatment.

42

5.1.2 Youths and therapists views of the alliance are not interchangeable.
Another aim of this thesis is to understand more of the relationship between youth and
therapist ratings, and to explore to what degree there is overlap in youths and therapists
implicit views of the alliance (paper II). The findings showed that associations between youth
and therapist perspectives were moderate, but investigations of the underlying factor structure
indicated that the two raters may perceive the alliance as having somewhat different
constructs. Analyses of the youth scale showed that a two-factor solution yielded the best
model fit, but that items clustered mainly based on valence i.e., whether the item was
positively or negatively worded and not by conceptual meaning i.e., whether the items
pertained to the task or bond dimension. This result was similar to the findings of Accurso et
al. (2013) and suggests that youth primarily distinguish between positive and negative aspects
of the alliance. In contrast, but consistent with the findings of Giuseppe et al. (1996), a factor
solution based on item content (i.e., task and bond) characterized the therapist ratings.
There may be several reasons why youth and therapists perceive the alliance
differently. For instance, the difference may be related to developmental level, as one might
expect increasing differentiation in concepts with maturation, including the alliance construct
(Shirk & Saiz, 1992). However, in a study by Accurso et al. (2013), caregiver ratings
clustered by valence and not by content. Furthermore, several studies have shown that adult
clients seem to perceive the alliance differently from their therapists (Andrusyna, Tang,
DeRubeis, & Luborsky, 2001; Bachelor, 2013; Horvath & Bedi, 2002), undermining the
assumption that the ability to distinguish the different conceptually-based dimensions is
related to maturity level alone. An alternative way of understanding this discrepancy is that
the perception of the alliance is related to degree of therapy experience and training.
Therapists have often met a variety of different families, making it easier to compare and
make more nuanced judgments of a particular child. In addition, many therapists may be
aware of the theoretical model of the therapeutic alliance and may thus be primed to view the
alliance along these dimensions. Since a mismatch in youth and therapist alliance ratings was
found to predict poorer outcome, it is important to further explore the implications that these
divergent conceptualizations may have for research and clinical practice.
5.1.3 Youths perceptions of caregiver approval of therapy predict dropout. A
novel finding in this thesis is that dropout was influenced by the youths perceptions of their
parents approval of the treatment (paper III). This finding is in line with our expectations and
corresponds with the theory of social referencing (see e.g., Morris et al., 2007). This theory
describes how children use their caregivers expressed emotions to guide their own
43

interpretations of a new and ambiguous situation as being safe or not. Since all of the youth in
our sample had experienced being unprotected and vulnerable during one or more traumatic
events, the youth would likely be more sensitive to and dependent on emotional support from
their caregivers. Because this study is the first to investigate treatment dropout in relation to
perceived parental support, we do not know whether such support is particularly important for
traumatized youth or whether the same findings pertain to other groups of children as well.
Another aspect of our findings was that age did not significantly moderate the effect of
youths perceived parental approval on dropout; adolescents as old as 18 still look to their
caregivers for emotional support. These results correspond with those of a study by Pinsof et
al. (2008), where adults treatment participation was influenced by whether they perceived
that their partner approved of the therapy. In sum, these results indicate that the need for
emotional support has no age limit, even if the need for practical support may decrease as one
grows older. Our results furthermore imply that the concept measured using the CAPPATS
scale may be useful and may help broaden our understanding of the relational processes that
contribute to treatment dropout, not only in traumatized populations but also in other
populations.
The finding that caregiver participation in the first session reduced the risk of dropout
was consistent with the results of another study investigating dropout among sexually abused
children (McPherson, Scribano, & Stevens, 2012). This aspect was significant even when
controlling for youths perceived parental support, indicating that caregivers practical
support is also important. This fact may imply that clinicians should encourage caregivers to
attend at least the first session in an attempt to reduce the risk of dropout. However, since the
caregivers were not randomized in their study participation, there may have been a selection
bias in which caregivers attended the first session. Consequently, we do not know whether it
was the actual participation itself that predicted dropout or whether participation was a
reflection of stronger family functioning. More information is required to reveal the role of
caregivers in the treatment process as providers of both practical and emotional support for
therapy attendance.
5.1.4 Linking the alliance to outcome: evaluating results in light of the pathways
model. A final aim of this thesis is to look at the main findings in light of the suggested
pathways model outlined in the introduction (Figure 1, p. 19) in order to see whether this can
help increase our understanding of how the alliance may be related to outcome. In this model,
the alliance is assumed to influence change in several ways. First, the establishment of an
initial alliance is assumed to be essential in order for youth and families to remain in
44

treatment and not drop out (Step 1). Subsequently, the alliance is hypothesized to produce
change through three different, but complementary, pathways. In the first pathway, a strong
alliance is assumed to be a healing factor in and of itself by providing a real relationship and a
corrective emotional experience. In the second pathway, the acts of agreeing on, and being
involving in, a set of therapeutic tasks is assumed to be beneficial by creating hope for
change. In this pathway, the expectations are themselves assumed to be the curative element,
and not the type of tasks per se. In contrast, in the third pathway, the alliance is assumed to
produce change by enhancing the involvement in specific health-promoting tasks, and it is
these tasks and assignments that are expected to be the real mechanisms of change.
Several findings in this thesis could be interpreted to be in line with the pathways
model. First, the link between an early alliance and treatment attendance was partially
supported, since at least therapist ratings of their relationship with the youth predicted
dropout. It was additionally expected that youth- and caregiver-rated alliances would predict
dropout, but this relationship might have been masked by methodological issues in our study
(see chapter 5.2.3).
Our findings furthermore suggest that the alliance was related to outcome through the
third pathway, i.e., involvement in specific and health-promoting tasks. For one, results
showed that the alliance scores were comparable across the two conditions, but it was only in
TF-CBT that the alliance predicted outcome. If the most important role of the alliance was to
promote hope through the agreement and involvement in any type of tasks (pathway 2), we
would have expected the alliance to predict outcome in both the treatment conditions.
Secondly, based on the similar levels of alliance ratings, it would be expected that the two
conditions would be equally beneficial. Instead, results from the TF-CBT study showed that
participants in the TF-CBT condition reported significantly lower symptom levels posttreatment (Jensen et al., 2014). Although the level of youths actual involvement in the
therapeutic tasks was not directly assessed in this thesis, the wording of the TASC-r asks the
youth to rate their active collaboration (e.g., I work with the therapist to make changes in my
life), indicating that the ratings may be close to actual involvement. However, it is a
weakness of this study that our findings cannot say what or which component(s) of TF-CBT
are active and beneficial. Based on the cognitive model of Ehlers & Clark (2000), it can be
assumed that the trauma narrative is an essential component. This finding is also in line with
results from Deblinger et al. (2011), but this question clearly warrants additional
investigation. It is important to note, however, that our finding does not show that the creation
of hope and expectations (pathway 2) is not important in the treatment of traumatized youth.
45

It was, for instance, found that the alliance predicted treatment satisfaction across both
treatment conditions, indicating that agreement and involvement, regardless of type of tasks,
may in some part be beneficial in and of itself.
Based on theory and the clinical literature, there is reason to assume that the first
pathway, where the alliance provides a real relationship and has the potential to act as a
corrective emotional experience, is important. For youth who have been traumatized within a
caretaking context, meeting a therapist who is able to understand their feelings and thoughts
and who is genuinely interested in their well being may be a new experience. This situation
may in turn help youth see themselves as people that deserve to be loved and be taken care of.
Furthermore, youth may be prompted to see other persons as benign and trustworthy. These
are core assumptions that may in turn be related to the reduction of PTSS and increased well
being. In this thesis it was not asked directly whether the youth perceived that their
relationship with the therapist provided such a corrective emotional experience, so the
relevance of this pathway cannot be directly evaluated. However, one can speculate as to
whether discrepancies in youth and therapist ratings of the alliance may have reflected a lack
of connectedness or a failure to attune to the youths experience, thereby resulting in poorer
outcomes.
Finally it should be considered to what degree the alliance-outcome relationships were
influenced by confounding factors. It has, for instance, been claimed that the alliance ratings
are merely a reflection of pretreatment characteristics such as attachment style or
interpersonal skills, and that these characteristics are the real predictors of outcome.
Furthermore, the ratings may be influenced by early symptom reduction, and the allianceoutcome relationship may be the result of the same rater halo effects (Crits-Cristoph et al.,
2006; DeRubeis et al., 2005). There are, however, several findings in this thesis that
strengthen the assumption that the alliance is at least in part a valid agent of therapeutic
change per se. First, if the alliance-outcome relationship was mostly a reflection of youths
background characteristics, we would have expected this relationship to manifest itself also in
the TAU condition since this trial was randomized. Second, in paper I we found that even
after controlling for early change, the alliance ratings at session 6 predicted subsequent
change in symptom level. Third, the youth-rated alliance was significantly related to clinicianrated outcome in paper I and marginally related to caregiver-rated outcome in paper II,
reducing the risk of same-rater halo effect. In sum, these findings strengthen the case of the
contribution of the therapeutic alliance.

46

5.1.5 The good news: the conditions for the alliance may be better than assumed.
The challenges facing therapists in forming a trusting and helpful relationship with patients
exposed to childhood trauma have been thoroughly described (J. A. Cohen et al., 2012; Paivio
& Patterson, 1999; Shirk & Eltz, 1998). Such challenges are particularly relevant in the
treatment of youth who have been exposed to interpersonal trauma in the caregiver
relationship, since these individuals may enter treatment with negative relationship
expectations. However, a general finding of this thesis was that the youth seemed to be
positively attuned to their therapist, even at the first session. Their initial alliance scores were
at the high end of the scale, significantly higher than those of the therapist. Moreover, the
youth ratings were consistent with the TASC-r scores reported in other clinical youth
populations (Accurso & Garland, 2014; Creed & Kendall, 2005; Kazdin et al., 2005). This
situation was the case even though the sample was multi-traumatized (experiencing, on
average, 3.6 different traumatic events) and the majority of the youth (59%) were exposed to
at least one traumatic event that occurred within the family context.
Another challenge is that therapists may fear that asking the youth about their
traumatic experiences will be upsetting and may undermine the therapeutic relationship
(Becker-Blease & Freyd, 2007; Hultmann, Mller, Ormhaug, & Broberg, 2014). However,
this assumption was not supported in this thesis. Audio recordings of the sessions showed that
there was substantially less focus on trauma talk or addressing trauma-related cognitions in
the TAU condition compared with the TF-CBT condition. Due to the youths impairing levels
of PTSS, thinking and talking about their traumatic events could be potentially painful and
trigger a significant amount of avoidance. As a result, working on the trauma narrative could
be a particularly challenging task. Even so, the mean alliance scores were comparable across
the two treatment conditions. Furthermore, even at session six, in most cases the time when
the exposure work had started, the therapeutic alliance ratings remained high in the TF-CBT
condition. This fact shows that most youth were not only able to tolerate the narrative work,
they also seemed to agree on this way of working and did not stop liking their therapist
although the tasks they were asked to perform were potentially difficult and anxiety
provoking.
In addition, although the TF-CBT therapists provided a treatment plan guided by a
manual, the therapeutic alliance was not undermined. This finding contrasts a commonly held
belief among many therapists (Addis & Krasnow, 2000; Nelson, Steele, & Mize, 2006), but is
consistent with previously reported results (Langer et al., 2011).

47

In sum, the findings of this thesis indicate that traumatized youth do not report lower
alliance rates compared with other diagnostic populations, and that neither work with a
trauma-focused, exposure-based treatment nor with a manual seem to threaten the building of
the alliance. Taken together, these findings are good news since studies have found that the
inclusion of an exposure-based component is important to reduce PTSS (Deblinger et al.,
2011; Ehlers et al., 2010), and the results from this thesis indicate that a strong alliance is an
important prerequisite for a good outcome using TF-CBT (paper I).

5.2 Methodological Considerations

5.2.1 The criterion validity of the therapeutic alliance scale. In order to yield valid
results, it is critical that the assessment procedures are able to capture the phenomenon that is
being studied. This refers to the criterion validity of a scale (Field, 2009). One critical caveat
in the youth literature is the lack of a common definition of the therapeutic alliance (Shirk et
al., 2011; Zack et al., 2007). This shortcoming influences fundamentally the alliance research,
since it raises the question of how the alliance can best be measured. In this thesis, the
alliance was defined as a two-dimensional construct consisting of emotional bond and
agreement on tasks (Shirk & Saiz, 1992). Little is known about whether agreement on goals
should be included as part of the youth alliance concept as well, or whether other dimensions
such as youths perceptions of parental treatment approval could substantially strengthen the
alliance concept in youth therapy. It is thus important to continue to assess the underlying
factor structure of the alliance scales in order to better understand how the alliance scales
should be constructed and interpreted.
5.2.2 The internal reliability of the questionnaires. Internal reliability refers to the
degree to which scores from tests or instruments are free from measurement errors (Pedhazur
& Schmelkin, 1991). This is commonly estimated by a scales internal consistency, often
reported in terPVRIWKH&URQEDFKVDOSKDLQGH[This term describes the relatedness
between items in one scale and, as such, indicates whether the items in a scale measure the
same construct. However, the term internal consistency may be somewhat misleading since
the alpha index is a measure not only of the magnitude of interrelatedness among items but
also the number of items included in the scale (Streiner, 2003). Although one cannot rely
solely on the alpha score alone when investigating the internal reliability of a scale, it is
commonly recommended that a scale have an alpha within the range of .70.90 (Tavakol &
Dennick, 2011). In our study, the majority of our scales had alpha values within the
recommended range of values. The only exception was the caregiver alliance scale, where the

48

internal reliability of the tDVNVXEVFDOHZDVLQDGHTXDWH $OWKRXJKRQO\WKHWRWDOVFRUH

ZDVXVHGLQWKLVVWXG\DQGWKHIXOOVFDOHKDGDQDGHTXDWHDOSKDOHYHO WKLVadequate
value could possibly be related to the increase in the number of items in the full scale (i.e., 6
vs. 12 items) (ibid.). It is therefore unclear to what degree the low reliability of one subscale
may have biased our findings. In this thesis, we failed to find the expected relationship
between caregiver alliance ratings and dropout, and it cannot be ruled out that this finding was
influenced by the potentially unfavorable psychometric properties of the scale.
5.2.3 Timing and source of ratings. In this thesis, the alliance was assessed after
sessions one and six, and all assessments were based on self-reports. There might, however,
have been better ways to capture the phenomenon of the alliance in youth therapy. For one,
with regard to timing of the assessment, it might be that session one was too early for youth
and caregivers to form a valid impression of their alliance with the therapist. In particular, the
degree to which youths and caregivers agreed on the therapeutic tasks may have been difficult
to evaluate. This fact could potentially explain why the internal reliability of the caregiver
task subscale was substantially lower than recommended. In contrast, since therapists have
often met a wide variety of youth during their careers it might be easier for therapists to form
more precise presumptions about their own abilities to form an alliance with a particular
youth or caregiver. Furthermore, the alliance was only assessed at two time points. Measuring
alliance trajectories is likely to provide a more nuanced way of evaluating associations
between therapy process and outcomes, and it has been suggested that at least four time points
should be included in order to fully understand the magnitude of the alliance-outcome
relationship (Crits-Cristoph et al., 2011).
Second, the use of self-reporting has its weaknesses. In particular, there is a risk that
the ratings may be biased by the respondents social desirability or self-desirability, which
refers to the tendency to present oneself in a good light to the researcher or interviewer
(Pedhazur & Schmelkin, 1991, p. 141). This aspect is particularly relevant to the alliance
ratings where youth and caregivers may have underreported potentially negative feelings
toward the therapist and the treatment in order not to hurt or upset the therapist. Although care
was taken to explain that the therapists would not be able to see their ratings, we still do not
know to what degree the alliance ratings may have been influenced. As an alternative,
observer ratings could have been used. These ratings have the advantage of being less
influenced by social desirability, and they reduce the burden of assessment on the treatment
participants (McLeod & Weisz, 2005). However, observed alliance have some disadvantages
too. For one, it does not capture the attitudinal and motivational aspects of the alliance (Elvins
49

& Green, 2008). Secondly, since the findings in this thesis indicate that youth may
conceptualize the alliance differently from therapists, it might be difficult for an outsider
observer to capture a youths personal experience of the alliance.
5.2.4 Attrition and missing data. Attrition is defined as the loss of participants after
randomization. Attrition can arise either because some participants decided to withdraw from
the study (treatment attrition) or because participants missed from one or more of the
scheduled assessments (measurement attrition). Although several measures were taken in our
study in order to follow up on the participants and reduce the number individuals who
dropped out from being included in the assessments, there were substantial levels of missing
data in all three studies. These missing data represent a potential threat to the validity of the
results for several reasons. For one reason, a loss of participants and assessments results in a
loss of power, which refers to the probability of correctly rejecting a false null hypothesis,
() usually interpreted as the probability of finding an effect when an effect exists.
(Shadish, Cook, & Campbell, 2002, p. 510). The second reason that missing data may be
problematic refers to the pattern of randomness in the missing data. Inspired by Rubins
article (1976), it is now common to classify missingness into three different patterns or
mechanisms: missing completely at random (MCAR), missing at random (MAR), or missing
not at random (MNAR). When there are no underlying variables related to the missingness
data are defined as MCAR. With such a pattern the missing data are ignorable and are not
assumed to bias the results. In the social sciences in general, and in clinical studies in
particular, missing data are seldom MCAR. Data are considered to be MAR if there is a
correlation between the missing data and other measured variables, but the fact that the data
are missing is not due to unobserved variables. Also, this pattern of missingness could be
ignored when correct analysis procedures are used. In contrast, we could not ignore missing
data if the missing mechanism is assumed to be MNAR. Such would be the case in our
sample if we have reason to expect that item non-responses or assessment attrition are related
to the symptoms measured.
Several steps were taken in order to evaluate the impact of these missing data and to
minimize how they bias our results. For instance, we ran the analyses with and without
multiply imputed data, and with and without weights. These methods are recommended
because they lead to less biased results compared with other simpler strategies, such as only
using complete case analyses with list-wise deletion (Graham, 2009; Schafer & Graham,
2002). In this thesis, we found that the results were similar, irrespective of which analytical
approaches were applied, including complete case analyses. This fact strengthens the validity
50

of our findings and implies that the missing data did not substantially skew our results.
However, no matter how sophisticated the missing data methods are, they can never replace
the value of a complete data set. As a result, future studies should continue to focus on how to
best obtain follow-up data on as many participants as possible.
5.2.5 The nested nature of the data. The data in our study had a hierarchical design
with youth nested within therapists and clinics. This arrangement introduces a certain degree
of dependence in our data, indicating that certain characteristics of individual therapists or
clinics might influence our findings. Consequently, we used statistical models that can
accommodate this dependency (West, 2009). One way to evaluate the degree of dependency
is to calculate the within-subjects intra-class correlations (ICC). The ICCs range in value from
0 to 1, with higher values indicating greater proportions of between-level variance and an
increased risk of possible bias if the nested nature of the data is not taken into account. If the
ICC values are small, multilevel models may be difficult or impossible to estimate, and in
practice they provide few benefits when the ICCs are below .05 (Dyer et al., 2005).
Investigations of the ICC levels in this thesis indicated that the effect of clinic was ignorable,
but that there was a small therapist effect (average ICC = .17). These results indicate that
multi-level analyses would be warranted, however efforts to take this nesting into account by
using multi-level factor analyses or linear mixed effects models produced unstable estimates.
This result could be due to the distribution of youth-therapist pairs in this sample, since 43.5%
of the therapists rated their alliance with one youth only, and only 30.7% of the therapists
rated their alliance with three or more youth. On average, therapists rated their alliance with
two youth (range: 16). Nevertheless, the use of single-level analyses in this thesis means that
a certain amount of bias in our findings cannot be ruled out.
5.2.6 The external validity of the findings. External validity refers to the degree to
which findings from one study can be applied to other samples that were not included in the
original study. Since we wanted to learn more about the treatment of traumatized youth under
regular conditions and therefore increase the generalizability of our findings, this study was
conducted in ordinary community clinics with referred patients and regular therapists with
normal caseloads. In addition, we purposefully adopted only a few exclusion criteria in order
to ensure that the recruited sample represented a wide range of children and caregivers.
However, there are still aspects of our study that could potentially challenge the external
validity of our results. For one, our sample was predominantly female (79.5%). Although it is
commonly found that girls report more PTSS and other related trauma reactions compared
with boys (Alisic et al., 2014; Tolin & Foa, 2008), and we would thus expect there to be more
51

girls than boys in a clinical sample of youth with PTSS, the fraction of females was still
higher compared with the general clinical population. According to national data, girls
comprise two thirds of the clinical population above 12 years old
(http://www.fhi.no/artikler/?id=84062). This fact means that we cannot rule out that there
might be some aspects of our findings that are less relevant to boys.
Another issue is that our sample was mostly composed of teenagers, which implies
that our results and conclusions might be less relevant to younger children. Although we
screened youth as young as 10, we found that only a few of the youngest individuals reported
3766DERYHWKHGHVLJQDWHGFXWRII&366VFRUHV
Furthermore, the decision to exclude families where there was a need for an interpreter
was made on the basis of reducing the potential confounds in our study. However, this choice
may have weakened the representativeness of our sample, since we know that several of the
traumatized youth in child and adolescent mental health clinics are refugees that have not yet
learned to speak Norwegian. It is thus less clear to what degree our findings are relevant to
individuals in this group.
Lastly, it is necessary to consider whether there are any systematic differences
between individuals who agreed to participate in the study and those who did not. Since
participation required that the caregivers and youth spend extra time to complete the
assessments and agree to have their sessions audiotaped, it might be that people who agreed to
be included in the study were more motivated for treatment and were less burdened by other
life stressors compared with people who declined to participate. This hypothesis could
partially explain why the alliance scores reported were higher than what were expected, and
why the dropout rates were lower than what has been reported in other studies (e.g.,
McPherson et al., 2012; Saxe, Ellis, Fogler, & Navalta, 2012). It would have been useful to
learn more about the individuals who did not wish to participate in terms of e.g., their levels
of trauma exposure and trauma symptoms. This information would shed light on the degree of
selection in our participating sample. Furthermore, there could also be a bias in the group of
therapists. Therapists who volunteered to participate in this study may have been more
motivated and engaged in their jobs than the general pool of therapists. This situation is
particularly relevant to TF-CBT therapists, since they volunteered to receive extra training
and may have been more motivated to learn and use new treatment methods. They also
received more supervision compared with the TAU therapists, an aspect that may have
influenced the quality of the treatment they provided and potentially biased our results. It is
unknown, however, to what degree this extra supervision may have influenced the alliance52

building since both youth and therapists reported similar alliance levels across the treatment
conditions. It can thus be assumed that although the supervision may have influenced the
comparison between the treatment procedures, this process may have had less of an impact on
the process and analyses of the alliance.
5.2.7 Interpretation of non-findings. It can be difficult to know how to interpret the
lack of a statistical relationship between two or more measured variables because absence of
evidence is not evidence of absence (Altmann & Bland, 1995, p. 485). If a relationship is not
statistically significant, it could be because of methodological aspects (such as low sample
size and lack of power), measurement issues, or other confounding factors that are masking an
existing relationship. It is therefore important to be cautious when drawing conclusions from a
non-finding. In this thesis, we failed to find several of the hypothesized relationships. We
expected, for instance, based on the youth alliance literature, to find a relationship between
the alliance and outcome in the TAU condition. This lack of a significant relationship could
be related to the timing of the outcome measurement, since maybe 15 sessions were not
enough time to capture the potential alliance effect. Another way of explaining our findings
involves the heterogeneity of methods used in the TAU condition. If the alliance works in
different ways in different treatment types, the aggregation of a variety of interventions could
mask potential associations. Finally, the lack of a significant relationship between caregiver
alliance and dropout must be considered. One way to understand this lack of relationship is
that there are other aspects besides the therapeutic relationship that weigh more heavily on
caregivers decisions about whether to continue or quit treatment, such as practical obstacles
or a feeling that the therapy is not relevant or is too demanding (Kazdin et al., 1997).
Alternatively, it might be that the therapeutic alliance is of importance from a caregiver
perspective, but that we attempted to measure its influence too early in our study. Another
methodological concern of ours was that the majority of caregivers reported scores at the high
end of the alliance scale, with low variance. It could be that this limited distribution made it
difficult for us to find a statistically significant association.

5.3 Clinical Implications

Several clinical implications can be drawn from the findings of this thesis. For one,
establishing a strong therapeutic alliance is important in order to help youth overcome their
PTSS. However, this alliance does not seem to be sufficient in and of itself, so therapists
should also include treatment strategies that are based on a theoretical rationale of how the
symptoms develop and are maintained. Conversely, introducing potentially active treatment

53

procedures may have little effect if these procedures are not presented in the context of a
warm and collaborative relationship.
The finding that youth and therapists seem to differ in their implicit views of the
alliance may have implications for therapists alliance-building strategies. Since our results
indicate that youth do not differentiate between the task and procedures in the therapy and
their emotional connection to the therapist, it may be that youth judge the relevance and
acceptance of a suggested task based on their liking or dislike of the therapist as a person.
However, it might also be that a lack of therapeutic tasks that make sense and seem relevant
to the youth may have a negative influence on the youths perceptions of his or her
relationship to the therapist. In practice, this viewpoint indicates that therapists should
incorporate therapeutic tasks as part of their alliance-building strategies since a solitary focus
on the emotional bond (building rapport) may not be sufficient in order to establish a strong
alliance.
Another implication of our work relates to the importance of including the youths
perspectives of their caregivers approval of the therapy. Investigating to what degree the
youth feel that they receive support from their parents for attending therapy, and potentially
addressing this issue with their caregivers, may be crucial in order to help youth remain in
therapy. Since the results of paper II showed that youths perceptions of parental approval
were not related to the caregivers own reports of their alliance to the therapist, youth may not
be very good at evaluating their caregivers actual attitudes toward the therapy; a clarification
of this aspect may in many cases be fruitful.
Furthermore, our results suggest that the therapeutic alliance and level of dropout were
not negatively affected in the TF-CBT condition compared with the TAU condition. These
results are encouraging since trauma-focused and exposure-based therapies are recommended
interventions for youth suffering from PTSS (AACAP, 2010; NICE, 2005). Clinically, this
finding implies that therapists might not need to fear that addressing youths traumatic
experiences will impede the formation of an alliance. At the same time, there was further no
evidence to suggest that youth will drop out more often from an exposure-based treatment
compared with an intervention with low levels of trauma focus.

5.4. Recommendations for Future Research

This thesis provides new and interesting results that open up possibilities for future
research. First, in order to better understand the pathways from the alliance to outcome, the
hypothesized link among the therapeutic alliance, involvement in treatment tasks, and

54

outcome (pathway 3) should be investigated. In particular, it will be interesting to investigate

whether the alliance can predict the youths level of involvement in the trauma narrative work
in TF-CBT and whether this involvement will again be related to outcome.
Another important aspect relates to how therapists proceed in establishing a strong
therapeutic alliance. This process is particularly important to youth for whom a close
relationship may be a trauma reminder and youth who are poorly motivated to receive
treatment. Furthermore, the finding that a low alliance agreement between therapists and
youths perspectives on the alliance seems to negatively influence the treatment process
implies that future studies should investigate how youth express their emotional liking and
agreement about therapeutic tasks. Such work could help therapists improve their clinical
judgments of the alliance. Additionally, learning more about which therapist behaviors
promote a strong alliance would improve therapist practices and provide effective therapy for
trauma-exposed youth.
In this thesis, we found that the alliance measured at one time point (session 6)
predicted post-treatment symptom reduction, but that alliance rated after session one were
unrelated to outcome (paper I). To date, little is known about how the alliance develops over
the course of treatment, and how this development relates to outcome. Future studies should
therefore aim to measure the alliance at additional time points. In the adult field, it has been
found that unrepaired ruptures of the alliance are related to poorer outcomes in PTSD
treatment (McLaughlin et al., 2013), underscoring the importance of learning more about how
therapists can establish and maintain a strong alliance in trauma treatment.
Furthermore, in order to more fully understand the concept of the alliance in youth
therapy, several aspects should be investigated. One of these aspects relates to the
dimensionality of the alliance. An ideal study would include several different alliance
measures so that the factor structure of these measures could be investigated with the same
sample. It would also be interesting to learn more about the potential role of the goal
dimension, since this subject has been rarely empirically studied in youth therapy. Another
issue relates to the systemic view of the alliance where the caregivers perspectives are
included in the alliance measures. In particular, youths perceptions of their caregivers views
hold promise as an important aspect, but additional studies are needed to more fully
understand how this interplay relates to the treatment process and outcome. It would also be
valuable to learn more about the caregivers actual impression of the therapy and to what
degree their approval is related to treatment engagement. Follow-up studies of the caregiver
alliance also seem warranted. Similarly, since the conceptualization of the alliance seems to
55

vary according to the source, therapist perspectives should also be included in future studies.
These perspectives may be related to different, but important, aspects of the therapy.

6. Conclusions
Overall, the results of this thesis show that the therapeutic alliance plays an important
role the treatment of traumatized youth. Specifically, a strong alliance was found to predict
symptom reduction (papers I and II), treatment satisfaction (paper II), and dropout (paper III).
However, the alliance-outcome relationship varied across treatment method and rater
perspective, and our findings suggest that there is also an interplay between the alliance
perspectives of different treatment participants. The results showed that the alliance-outcome
relationship was moderated by treatment method in that a positive working relationship
appeared to be particularly important in the TF-CBT intervention but not in TAU.
Furthermore, youth and therapist ratings of the alliance were differentially related to outcome,
and additional information was gained when therapist ratings were seen in relation to the
youths perspectives. Last, aspects of the youth-caregiver relationship were found to influence
the treatment process, indicating that a systemic perspective on the alliance is warranted in
order to more fully understand how this relationship variable interacts with the treatment
process. Taken together, these findings indicate that although the therapeutic alliance is
present in all types of treatments, in the sense that no treatments are carried out without a
relational context (Safran & Muran, 2000), the way the therapeutic alliance is related to
outcome and how it is conceptualized may depend on the treatment type and source of rating.
This contextualization of the alliance could help explain the wide range of alliance-outcome
correlations reported in different studies (McLeod, 2011; Shirk et al., 2011), and the range of
methodological moderators influencing this relationship (McLeod, 2011). The results imply
that direct comparisons of the alliance-outcome relationship across different studies may be
less fruitful and focus should remain on understanding more how the alliance is embedded in
the treatment context. Additionally, our findings suggest that investigations of specific
methods should be viewed within the relational context the components are provided.
Finally, although the results in this thesis may help elucidate the pathways from
alliance to outcome, further investigations should be conducted to strengthen our
understanding of this relationship and improve the treatments provided to traumatized youths
and their caregivers.

56

References
AACAP. (2010). Practice parameter for the assessment and treatment of children and
adolescents with posttraumatic stress disorder|. Journal of the American Academy of
Child and Adolesecent Psychiatry, 49(4), 414-430.
Accurso, E. C., & Garland, A. F. (2014). Child, caregiver, and therapist perspectives on the
therapeutic alliance in usual care child psychotherapy. Psychological Assessment. doi:
10.1037/pas0000031
Accurso, E. C., Hawley, K. M., & Garland, A. F. (2013). Psychometric properties of the
Therapeutic Alliance Scale for Caregivers and Parents. Psychological Assessment,
25(1), 244-252. doi: 10.1037/a0030551
Addis, M. E., & Krasnow, A. D. (2000). A national survey of practicing psychologists
attitudes toward psychotherapy treatment manuals. Journal of Consulting and Clinical
Psychology, 68, 331-339. doi: 10.1037/0022-006X.68.2.331
Alisic, E., Zalta, A. K., van Wessel, F., Larsen, S. E., Hafstad, G. S., Hassanpour, K., & Smid,
G. E. (2014). Rates of post-traumatic stress disorder in trauma-exposed children and
adolescent: meta-analysis. British Journal of Psychiatry, 204, 335-340. doi:
10.1192/bjp.bp.113.131227
Altmann, D. G., & Bland, J. M. (1995). Statistics notes. Absence of evidence is not evidence
of absence. British Medical Journal, 311, 485.
Anda, R. F., Felitti, V. J., Bremner, J. D., Walker, J. D., Whitfield, C., Perry, B. D., . . . Giles,
W. H. (2006). The enduring effects of abuse and related adverse experiences in
childhood. European Archives if Psychiatry and Clinical Neuroscience, 256(3), 174186. doi: 10.1007/s00406-005-0624-4
Andrusyna, T. P., Tang, T. Z., DeRubeis, R. J., & Luborsky, L. (2001). The factor structure of
the Working Alliance Inventory in cognitive-behavioral therapy. The Journal of
Psychotherapy Practice and Research, 10, 173-178.
Angold, A., Costello, E. J., Messer, S. C., & Pickles, A. (1995). Development of a short
questionnaire for use in epidemioligical studies of depression in children and
adolescents. International Journal of Methods in Psychiatric Research, 5, 237-249.
APA. (2000). Diagnostic and statistical manual of mental disorders (4th ed., text revision)
(DSM-IV-TR). Washington, DC: American Psychiatric Association.

APA. (2013). Diagnostic and Statistical Manual of Mental Disorders (5th Ed.) (DSM-5).
Arlington, VA: American Psychiatric Publishing.
Arnow, B. A., Steidtman, D., Blasey, C., Manber, R., Constantino, M. J., Klein, D. N., . . .
Kocsis, J. H. (2013). The relationship between the therapeutic alliance and treatment
outcome in two distinct psychotherapies for chronic depression. Journal of Consulting
and Clinical Psychology, 81(4), 627-638. doi: 10.1037/a0031530
Asnaani, A., & Foa, E. B. (2014). Expanding thet lens of evidence-based practice in
psychotherapy to include a common factors perspective: Comment on Laska, Gurman,
and Wampold. Psychoterapy, 51(4), 487-490. doi: 10.1037/a0036891
Bachelor, A. (2013). Client's and therapists' views of the therapeutic alliance: similarities,
differences and relationship to therapy outcome. Clinical Psychology and
Psychotherapy, 20, 118-135. doi: 10.1002/cpp.792
Barendse, M. T., Oort, F. J., & Timmerman, M. E. (2015). Using exploratory factor analyses
to determine the dimensionality of discrete responses. Structural Equation Modeling:
A Multidisciplinary Journal, 22(1), 87-101. doi: 10.1080/10705511.2014.934850
Becker-Blease, K. A., & Freyd, J. J. (2007). The ethics of asking about abuse and the harm of
"don't ask, don't tell". American Psychologist, 62(4), 330-332. doi: 10.1037/0003066X.62.4.330
Benish, S. G., Imel, Z. E., & Wampold, B. E. (2008). The relative efficacy of bona fide
psychotherapies for treating post-traumatic stress disorder: a meta-analysis of direct
comparisons. Clinical Psychology Review, 28, 746-758.
Birmaher, B., Brent, D. A., Chiappetta, L., Bridge, J., Monga, S., & Baugher, M. (1999).
Psychometric properties of the Screen for Child Anxiety Related Emotional Disorders
(SCARED): a replication study. Journal of the American Academy of Child and
Adolesecent Psychiatry, 38(10), 1230-1236. doi: 10.1097/00004583-19991000000011
Bordin, E. S. (1979). The generalizability of the psychoanalytic concept of the working
alliance. Psychotherapy: Theory, Research & Practice, 16(3), 252-260.
doi: http://dx.doi.org/10.1037/h0085885
Bowlby, J. (1988). A secure base: Clinical applications of attachment theory. London:
Tavistock/Routledge.
Brookman-Frazee, L., Haine, R. A., Gabayan, E. N., & Garland, A. F. (2008). Predicting
frequency of treatment visits in community-based youth psychotherapy. Psychological
Services, 5(2), 126-138. doi: 10.1037/1541-1559.5.2.126
58

Campos, J. J. (1984). A new perspective on emotions. Child Abuse & Neglect, 8(2), 147-156.
Carr, A. (2004). Interventions for post-traumatic stress disorder in children and adolescents.
Pediatric Rehabilitation, 7(4), 231-244. doi: 10.1080/13638490410001727464
Chasson, G. S., Vincent, J. P., & Harris, G. E. (2008). The use of symptom severity measured
just before termination to predict child treatment dropout. Journal of Clinical
Psychology, 64(7), 891-904. doi: 10.1002/jclp.20494
Clark, L. A., & Watson, D. (1995). Constructing validity: basic issues in objective scale
development. Psychological Assessment, 7(3), 309-319.
Cloitre, M., Cohen, L. R., & Scarvalone, P. (2002). Understanding revictimization among
childhood sexual abuse survivors: An interpersonal schema approach. Journal of
Cognitive Psychotherapy: An Internatinonal Quarterly, 16(1), 91-111.
Cloitre, M., Koenen, K. C., Cohen, L. R., & Han, H. (2002). Skills training in affective and
interpersonal regulation followed by exposure: A phase based treatment for PTSD
related to childhood abuse. Journal of Consulting and Clinical Psychology, 70(5),
1067-1074. doi: 10.1037/0022-006X.70.5.1067
Cloitre, M., Stovall-McClough, K. C., Miranda, R., & Chemtob, C. M. (2004). Therapeutic
alliance, negative mood regulation, and treatment outcome in child abuse-related
posttraumatic stress disorder. Journal of Consulting and Clinical Psychology, 72(3),
411-416. doi: 10.1037/0022-006X.72.3.411
Cohen, J. (1992). A power primer. Psychological Bulletin, 112, 155-159. doi: 10.1037/00332909.112.1.155
Cohen, J. A., Deblinger, E., Mannarino, A. P., & Steer, R. A. (2004). A multisite, randomized
controlled trial for children with sexual abuse-related PTSD symptoms. Journal of the
American Academy of Child & Adolescent Psychiatry, 43(4), 393-402.
doi: http://dx.doi.org/10.1097/00004583-200404000-00005
Cohen, J. A., & Mannarino, A. P. (1996). A treatment outcome study for sexually abused
preschool children: initial findings. Journal of American Academy of Child and
Adolescent Psychiatry, 35(1), 42-50.
Cohen, J. A., & Mannarino, A. P. (1998). Interventions for sexually abused children: initial
treatment outcome findings. Child Maltreatment, 3, 17-26.
Cohen, J. A., Mannarino, A. P., & Deblinger, E. (2006). Treating trauma and traumatic grief
in children and adolescents. New York, NY: Guilford Press; US.

59

Cohen, J. A., Mannarino, A. P., & Iyengar, S. (2011). Community treatment of posttraumatic
stress disorder for children exposed to intimate partner violence: A randomized
controlled trial. Archives of Pediatrics Adolescent Mental Health, 165(1), 15-21.
Cohen, J. A., Mannarino, A. P., Kliethermes, M., & Murray, L. A. (2012). Trauma-focuced
CBT for youth with complex trauma. Child Abuse & Neglect, 36, 528-541. doi:
10.1016/j.chiabu.2012.03.007
Cohen, J. A., Mannarino, A. P., & Kundsen, K. (2005). Treating sexually abused children: 1
year follow-up of a randomized controlled trial. Child Abuse & Neglect, 29, 135-145.
Cohen, J. A., Mannarino, A. P., Perel, J. M., & Staron, V. (2007). A pilot randomized
controlled trial of combined trauma-focused CBT and sertraline for childhood PTSD
symptoms. Journal of the American Academy of Child and Adolesecent Psychiatry,
46, 811-819. doi: 10.1097=chi.0b013e3180547105
Cook, A., Spinazzola, J., Ford, J. D., Lanktree, C., Blaustein, M., Cloitre, M., . . . van der
Kolk, B. (2005). Complex trauma in children and adolescents. Paychiatric Annals,
35(5), 390-398.
Copeland, W. E., Keeler, G., Angold, A., & Costello, E. J. (2007). Traumatic events and
posttraumatic stress in childhood. Archives of General Psychiatry, 64, 577-584.
Creed, T. A., & Kendall, P. C. (2005). Therapist alliance-building behavior with a cognitivebehavioral treatment for anxiety in youth. Journal of Consulting and Clinical
Psychology, 73(3), 498-505. doi: 10.1037/0022-006X.73.3.498
Crits-Cristoph, P., Gibbons, M. B. C., Hamilton, J., Ring-Kurtz, S., & Gallop, R. (2011). The
dependability of alliance assessments: The alliance - outcome correlation is larger than
you might think. Journal of Consulting and Clinical Psychology, 79(3), 267-278. doi:
10.1037/a0023668
Crits-Cristoph, P., Gibbons, M. B. C., & Hearon, B. (2006). Does the alliance cause good
outcome? Recommendations for future research on the alliance. Psychotherapy:
Theory, Research & Practice, 43(3), 280-285. doi: 10.1037/0033-3204.43.3.280
Cummings, C. M., Caporino, N. E., Settipani, C. A., Read, K. L., Compton, S. N., March, J., .
. . Kendall, P. C. (2013). The therapeutic relationship in cognitive-behavioral therapy
and pharmacotherapy for anxious youth. Journal of Consulting and Clinical
Psychology, 81(5), 859-864. doi: 10.1037/a0033294
Davies, M. G. (1995). Parental distress and ability to cope following disclosure of extrafamilial sexual abuse. Child Abuse & Neglect, 19, 399-408.

60

Deblinger, E., Lippman, J., & Steer, R. A. (1996). Sexually abused children suffering
posttraumatic stress symptoms: Initial treatment outcome findings. Child
Maltreatment, 1, 310-321. doi: 10.1177/1077559596001004003
Deblinger, E., Mannarino, A. P., Cohen, J. A., Runyon, M. K., & Steer, R. A. (2011).
Trauma-focused cognitive behavioral therapy for chidlren: Impact of the trauma
narrative and treatment length. Depression and Anxiety, 28, 67-75. doi:
10.1002/da.20744
Deblinger, E., Mannarino, A. P., Cohen, J. A., & Steer, R. A. (2006). A Follow-up Study of a
Multisite, Randomized, Controlled Trial for Children With Sexual Abuse-Related
PTSD Symptoms. Journal of the American Academy of Child & Adolescent
Psychiatry, 45(12), 1474-1484.
doi: http://dx.doi.org/10.1097/01.chi.0000240839.56114.bb
Deblinger, E., Stauffer, L., & Steer, R. A. (2001). Comparative efficacies of supportive and
cognitive behavioral group therapies for young children who have been sexually
abused and their nonoffending mothers. Child Maltreatment, 6, 332-343.
Del Re, A. C., Flckiger, C., Horvath, A. O., Symonds, D., & Wampold, B. E. (2012).
Therapist effects in the therapeutic alliance-outcome relationship: A restrictedmaximum likelihood meta-analysis. Clinical Psychology Review, 32(7), 642-649. doi:
10.1016/j.cpr.2012.07.002
DePrince, A. P., Combs, M. D., & Shanahan, M. (2009). Automatic relationship-harm
associations and interpersonal trauma involving close others. Psychology of Women
Quarterly, 33, 163-171.
DeRubeis, R. J., Brotman, M. A., & Gibbons, C. J. (2005). A conceptual and methodological
analysis of the nonspecifics argument. Clinical Psychology Science and Practice, 12,
174-183. doi: 10.1093/clipsy/bpi022
DiGiuseppe, R., Linscott, J., & Jilton, R. (1996). Developing the therapeutic alliance in childadolescent psychotherapy. Applied & Preventive Psychology, 5, 85-100.
Dube, S. R., Felitti, V. J., Dong, M., Giles, W. H., & Anda, R. F. (2003). The impact of
adverse childhood experiences on health problems: evidence from four birth cohorts
dating back to 1900. Preventive Medicine, 37, 268-277. doi: 10.1016/S00917435(03)00123-3
Dyb, G. (2007). Ta barnet p alvor - ogs i forskning. Forskningsetikk, 1, 12-15.

61

Dyer, N. G., Hanges, P. J., & Hall, R. J. (2005). Applying multilevel confirmatory factor
analysis techniques to the study of leadership. The Leadership Quarterly, 16, 149-167.
doi: 10.1016/j.leaqua.2004.09.009
Ehlers, A., Bisson, J., Clark, D. M., Creamer, M., Pilling, S., Richards, D., . . . Yule, W.
(2010). Do all psychological treatments really work the same in posttraumatic stress
disorder? Clinical Psychology Review, 30, 269-276. doi: 10.1016/j.cpr.2009.12.001
Ehlers, A., & Clark, D. M. (2000). A cognitive model of posttraumatic stress disorder.
Behaviour Research and Therapy, 38, 319-345.
Ehring, T., Welboren, R., Morina, N., Wicherts, J. M., Freitag, J., & Emmelkamp, P. M. G.
(2014). Meta-analysis of psychological treatment for posttraumatic stress disorder of
adult survivors of childhood abuse. Clinical Psychology Review, 34, 645-657. doi:
10.1016/j.cpr.2014.10.004
Elhai, J. D., Miller, M. E., Ford, J. D., Biehn, T. L., Palmieri, P. A., & Frueh, B. C. (2012).
Posttraumatic stress disorder in DSM-5: Estimates of prevalence and symptom
structure in a nonclinical sample of college students. Journal of Anxiety Disorders,
26(1), 58-64. doi: 10.1016/j.janxdis.2011.08.013
Elliot, A. N., & Carnes, C. N. (2001). Reactions of nonoffending parents to the sexual abuse
of their child: a review of the literature. Child Maltreatment, 6, 314-331.
Eltz, M. J., Shirk, S. R., & Sarlin, N. (1995). Alliance formation and treatment outcome
among maltreated adolescents. Child Abuse & Neglect, 19(4), 419-431.
doi: http://dx.doi.org/10.1016/0145-2134%2895%2900008-V
Elvins, R., & Green, J. (2008). The conceptualization and measurement of therapeutic
alliance: An empirical review. Clinical Psychology Review, 28, 1167-1187. doi:
10.1016/j.cpr.2008.04.002
Eslinger, J. G., Sprang, G., & Otis, M. D. (2012). Child and caregiver dropout in child
psychotherapy for trauma. Journal of Loss and Trauma: International Perspectives on
Stress & Coping. doi: DOI:10.1080/15325024.2012.742720
Faw, L., Hogue, A., Johnson, S., Diamond, G. M., & Liddle, H. A. (2005). The Adolescent
Therapeutic Alliance Scale (ATAS): Initial psychometrics and prediction of outcome
in family-based substance abuse prevention counseling. Psychoterapy Research, 15(12), 141-154. doi: 10.1080/10503300512331326994
Feinman, S. (1992). Social referencing and the social construction of reality in infancy. New
York, NY: Plenum Press.
Field, A. P. (2009). Discovering statistics using SPSS. London: Sage.
62

Finkelhor, D., Ormrod, R. K., & Turner, H. A. (2007). Polyvictimization and trauma in a
national longitunial cohort. Development and Psychopathology, 19, 149-166. doi:
10.10170S0954579407070083
Fjermestad, K., McLeod, B. D., Heiervang, E. R., Havik, O. E., st, L.-G., & Haugland, B. S.
M. (2012). Factor structure and validity of the Therapy Process Observational Coding
System for child psychotherapy - Alliance scale. Journal of Clinical Child &
Adolescent Psychology, 41(2), 246-254. doi: 10.1080/15374416.2012.651999
Flckiger, C., Del Re, A. C., Wampold, B. E., Symonds, D., & Horvath, A. O. (2012). How
central is the alliance in psychotherapy? A multilevel longitudinal meta-analysis.
Journal of Counceling Psychology, 59(1), 10-17. doi: 10.1037/a0025749
Foa, E. B., Johnson, K. M., Feeny, N. C., & Treadwell, K. R. (2001). The Child PTSD
Symptom Scale: A preliminary examination of its psychometric properties. Journal of
Clinical Child Psychology, 30(3), 376-384.
doi: http://dx.doi.org/10.1207/S15374424JCCP3003_9
Furman, W., Simon, V. A., Shaffer, L., & Bouchery, H. A. (2002). Adolescent's working
models and styles for relationshops with parents, friends, and romantic partners. Child
Development, 73(1), 241-255.
Garcia, J. A., & Weisz, J. R. (2002). When youth mental health care stops: Therapeutic
relationship problems and other reasons for ending youth outpatient treatment. Journal
of Consulting and Clinical Psychology, 70(2), 439-443. doi: 10.1037//0022006X.70.2.439
Garland, A. F., Aarons, G. A., Hawley, K. M., & Hough, R. L. (2003). Relationship of youth
satisfaction with mental health services and changes in symptoms and functioning.
Psychiatric Services, 54(11), 1544-1546.
Garland, A. F., Haine-Schlagel, R., Accurso, E. C., & Baker-Ericzn, M. J. (2012). Exploring
the effect of therapists' treatment practices on client attendance in community-based
care for children. Psychological Services, 9(1), 74-88. doi: 10.1037/a0027098
Gelso, C. J. (2009). The real relationship in a postmodern world: Theoretical and empirical
explorations. Psychotherapy Research, 19(3), 253-264. doi:
10.1080/10503300802389242
Gerson, R., & Rappaport, N. (2013). Traumatic stress and posttraumatic stress disorder in
youth: Resent research findings on clinical impact, assessment, and treatment. Journal
of Adolescent Health, 25, 137-143. doi: 10.1016/j.jadohealth.2012.06.018

63

Goodman, R. (2001). Psychometric properties of the Strengths and Difficulties Questionnaire.

Journal of the American Academy of Child and Adolesecent Psychiatry, 40(11), 13371345.
Gopalan, F., Goldstein, L., Klingenstein, K., Sicher, C., Blake, C., & McKay, M. M. (2010).
Engaging families into child mental health treatment: updates and special
considerations. Journal of Canadian Academy of Child and Adolescent Psychiatry,
19(3), 182-196.
Graham, J. W. (2009). Missing data analysis: Making it work in the real world. Annual
Review of Psychology, 60, 549-576. doi: 10.1146/annurev.psych.58.110405.085530
Green, J. (2006). Annotation: The therapeutic alliance a significant but neglected variable in
child mental health treatment studies. Journal of Child Psychology and Psychiatry,
47(5), 425-435. doi: 10.1111/j.1469-7610.2005.01516.x
Griffin, M. G., Resick, P. A., Waldrop, A. E., & Mechanic, M. B. (2003). Participation in
trauma research: Is there evidence of harm? Journal of Traumatic Stress, 16(3), 221227.
Hafstad, G. S., Dyb, G., Jensen, T. K., Steinberg, A. M., & Pynoos, R. S. (2014). PTSD
prevalence and symptom structure of DSM-5 criteria in adolescents and young adults
surviving the 2011 shooting in Norway. Journal of Affective Disorders, 169, 40-46.
doi: 10.1016/j.jad.2014.06.055
Hartman, D., & Zimberoff, D. (2004). Corrective emotional experience in the therapeutic
process. Journal of Heart-Centered Therapies, 7(2), 3-84.
Hawley, K. M., & Garland, A. F. (2008). Working alliance in adolescent outpatient therapy:
Youth, parent and therapist reports and associations with therapy outcomes. Child &
Youth Care Forum, 37(2), 59-74. doi: http://dx.doi.org/10.1007/s10566-008-9050-x
Hawley, K. M., & Weisz, J. R. (2005). Youth versus parent working alliance in usual clinical
care: Distinctive associations with retention, satisfaction, and treatment outcome.
Journal of Clinical Child & Adolescent Psychology, 34(1), 117-128. doi:
10.1207/s15374424jccp3401_11
Hersoug, A. G., Hglend, P., Havik, O. E., von der Lippe, A., & Monsen, J. T. (2009).
Pretreatment patient characteristics related to the level and development of working
alliance in long-term psychotherapy. Psychoterapy Research, 19(2), 172-180. doi:
10.1080/10503300802657374
Hogue, A., Dauber, S., Stambaugh, L. F., Cecero, J. J., & Liddle, H. A. (2006). Early
therapeutic alliance and treatment outcome in individual and family therapy for
64

adolescent behavior problems. Journal of Consulting and Clinical Psychology, 74(1),

121-129. doi: 10.1037/0022-006x.74.1.121
Holt, T., Cohen, J. A., Mannarino, A. P., & Jensen, T. K. (2014). Parental emotional response
to children's traumas. Journal of Aggression, Maltreatment, & Trauma, 1057-1071.
doi: 10.1080/10926771.2014.953717
Holt, T., Jensen, T. K., & Wentzel-Larsen, T. (2014). The change and the mediating role of
parental emotional reactions and depression in the treatment of traumatized youth:
results from a
randomized controlled study. Child and Adolescent Psychiatry & Mental Health, 8(11).
Horvath, A. O., & Bedi, R. P. (2002). The alliance. In J. C. Norcross (Ed.), Psychotherapy
Relationships That Work: Therapist Contributions and Responsiveness to Patients
(pp. 37-70). New York: Oxford University Press.
Horvath, A. O., Del Re, A. C., Flckiger, C., & Symonds, D. (2011). Alliance in individual
psychotherapy. Psychoterapy, 48(1), 9-16. doi: 10.1037/a0022186
Hougaard, E. (1994). The therapeutic alliance: A conceptual analysis. Scandinavian Journal
of Psychology, 35(1).
Hultmann, O., Mller, J., Ormhaug, S. M., & Broberg, A. (2014). Asking routinely about
intimate partner violence in a child and adolescent psychiatric clinic: a qualitative
study. Journal of Family Violence, 29, 67-78. doi: 10.1007/s10896-013-9554-5
Janoff-Bulman, R. (1992). Shattered assumptions: Towards a new psychology of trauma.
New York: Free Press.
Jaycox, L. H., Cohen, J. A., Mannarino, A. P., Walker, J. D., Langley, A. K., Gegenheimer,
K. L., . . . Schonlau, M. (2010). Childrens mental health care following Hurricane
Katrina: A field trial of trauma-focused psychotherapies. Journal of Traumatic Stress,
23. doi: 10.1002/jts.20518
Jensen, T. K., Haavind, H., Gulbrandsen, W., Mossige, S., Reichelt, S., & Tjersland, O. A.
(2010). What constitutes a good working alliance in therapy with children that may
have been sexually abused? Qualitative Social Work. doi:
10.1177/1473325010374146
Jensen, T. K., Holt, T., Ormhaug, S. M., Egeland, K., Granly, L. B., Hoaas, L. C., . . .
Wentzel-Larsen, T. (2014). A randomized effectiveness study comparing TraumaFocused Cognitive Behavioral Therapy with therapy as usual for youth. Journal of
Clinical Child & Adolescent Psychology, 43(3), 356-369. doi:
10.1080/15374416.2013.822307
65

Johnson, L. N., Ketring, S. A., & Anderson, S. R. (2013). Confirmatory factor analysis of the
family therapy alliance scale: An evaluation of factor structure across parents and
adolescents referred for At Risk Services. Contemporary Family Therapy, 35, 121136. doi: 10.1007/s10591-012-9221-7
Karver, M. S., Handelsman, J. B., Fields, S., & Bickman, L. (2005). A theoretical model of
common process factors in youth and family therapy. Mental Health Services
Research, 7(1), 35-51. doi: 10.1007/s11020-005-1964-4
Kassam-Adams, N., & Newman, E. (2005). Child and parent reactions to participation in
clinical research. General Hospital Psychiatry, 27, 29-35. doi:
10.1016/j.genhosppsych.2004.08.007
Kazdin, A. E. (2009). Understanding how and why psychotherapy leads to change.
Psychotherapy Research, 19(4-5), 418-428. doi: 10.1080/10503300802448899
Kazdin, A. E., Holland, L., & Crowley, M. (1997). Family experience of barriers to treatment
and premature termination from child therapy. Journal of Consulting and Clinical
Psychology, 65(3), 453-463.
Kazdin, A. E., Marciano, P. L., & Whitley, M. K. (2005). The therapeutic alliance in
cognitivebehavioral treatment of children referred for oppositional, aggressive, and
antisocial behavior. Journal of Consulting and Clinical Psychology, 73(4), 726-730.
doi: 10.1037/0022-006X.73.4.726
Kazdin, A. E., & Nock, M. K. (2003). Delineating mechanisms of change in child and
adolescent therapy: methodological issues and research recommendations. Journal of
Child Psychology and Psychiatry, 44(8), 1116-1129. doi: 10.1111/1469-7610.00195
Kazdin, A. E., & Whitley, M. K. (2006). Pretreatment social relations, therapeutic alliance,
and improvements in parenting practices in Parent Management Training. Journal of
Consulting and Clinical Psychology, 74(2), 345-355. doi: 10.1037/0022006X.74.2.346
Kazdin, A. E., Whitley, M. K., & Marciano, P. L. (2006). Childtherapist and parent
therapist alliance and therapeutic change in the treatment of children referred for
oppositional, aggressive, and antisocial behavior. Journal of Child Psychology and
Psychiatry, 47(5), 436-445. doi: 10.1111/j.1469-7610.2005.01475.x
Kearney, C. A., Wechsler, A., Kaur, H., & Lemos-Miller, A. (2010). Posttraumatic stress
disorder in maltreated youth: A review of contemporary research and thought. Clinical
Child and Family Psychology Review, 13, 46-76. doi: 10.1007/s10567-009-0061-4

66

Keller, S. M., Zoellner, L. A., & Feeny, N. C. (2010). Understanding factors associated with
early therapeutic alliance in PTSD treatment: adherence, childhood sexual abuse
history, and social support. Journal of Consulting and Clinical Psychology, 78(6),
974-979. doi: 10.1037/a0020758
Kelley, S. J. (1990). Parental stress response to sexual abuse and ritualistic abuse of children
in day-care centres. Nursing Research, 39(25-29).
Kendall, P. C., Comer, J. S., Marker, C. D., Creed, T. A., Puliafico, A. C., Hughes, A. A., . . .
Suveg, C. (2009). In-session exposure tasks and therapeutic alliance across the
treatment of childhood anxiety disorders. Journal of Consulting and Clinical
Psychology, 77(3), 517-525. doi: 10.1037/a0013686
Kilpatrick, D. G., Resnick, H. S., Milanak, M. E., Miller, M. W., Keyes, K. M., & Friedman,
M. J. (2013). National estimates of exposure to traumatic events and PTSD prevalence
using DSM-IV and DSM-5 criteria. Journal of Traumatic Stress, 26, 537-547. doi:
10.1002/jts.21848
Kilpatrick, D. G., Ruggiero, K. J., Acierno, R., Saunders, B. E., Resnick, H. S., & Best, C. L.
(2003). Violence and risk of PTSD, major depression, substance abuse/dependence,
and comorbidity: results from the national survey of adolescents. Journal of
Consulting and Clinical Psychology, 71(4), 692-700. doi: 10.1037/0022006X.71.4.692
King, N. J., Tonge, B. J., Mullen, P., Myerson, N., Heyne, D., Rollings, S., . . . Ollendick, T.
H. (2000). Treating sexually abused children with posttraumatic stress symptoms: A
randomized clinical trial. Journal of the American Academy of Child and Adolesecent
Psychiatry, 39(11), 1347-1355. doi: 10.1097/00004583-200011000-00008
Langer, D. A., McLeod, B. D., & Weisz, J. R. (2011). Do treatment manuals undermine
youth-therapist alliance in community clinical practice? Journal of Consulting and
Clinical Psychology, 79(4), 427-432. doi: 10.1037/a0023821
Lau, A. S., & Weisz, J. R. (2003). Reported maltreatment among clinic-referred children:
implications for presenting problems, treatment attrition, and long-term outcomes.
Journal of American Academy of Child and Adolescent Psychiatry, 42(11), 13271334. doi: 10.1097/01.CHI.0000085754.71002.14
Lawson, D. M. (2009). Understanding and treating children who experience interpersonal
maltreatment: empirical findings. Journal of Counceling and Development, 87, 204215.

67

Levin, L., Henderson, H. A., & Ehrenreich-May, J. (2012). Interpersonal predictors of early
therapeutic alliance in a transdiagnostic cognitive-behavioral treatment for adoelscents
with anxiety and depression. Psychotherapy, 49(2), 218-230. doi: 10.1037/a0028265
Marcus, D. K., O'Connell, D., Norris, A. L., & Sawaqdeh, A. (2014). Is the Dodo bird
endangered in the 21st century? A meta-analysis of treatment comparisons studies.
Clinical Psychology Review. doi: 10.1016/j.cpr.2014.08.001
Marker, C. D., Comer, J. S., Abramova, V., & Kendall, P. C. (2013). The reciprocal
relationship between alliance and symptom improvement across the treatment of
childhood anxiety. Journal of Clinical Child & Adolescent Psychology, 42(1), 22-33.
doi: 10.1080/15374416.2012.723261
Martin, D. J., Garske, J. P., & Davis, K. M. (2000). Relation of the therapeutic alliance with
outcome and other variables: a meta-analytic review. Journal of Consulting and
Clinical Psychology, 68(3), 438-450. doi: 10.1037/0022-006x.68.3.438
McGloin, J. M., & Widom, C. S. (2001). Resilience among abused and neglected children
grown up. Development and Psychopathology, 13, 1021-1038.
McKay, M. M., Lynn, C. J., & Bannon, W. M. (2005). Understanding inner city child mental
health need and trauma exposure: Implications for preparing urban service providers.
American Journal of Orthopsychiatry, 75(2), 201-210. doi: 10.1037/00029432.75.2.201
McLaughlin, A. A., Keller, S. M., Feeny, N. C., Youngstrom, E. A., & Zoellner, L. A. (2013).
Patterns of therapeutic alliance: rupture-repair episodes in prolonged exposure for
posttraumatic stress disorder. Journal of Consulting and Clinical Psychology. doi:
10.1037/a0034696
McLeod, B. D. (2011). Relation of the alliance with outcomes in youth psychotherapy: A
meta-analysis. Clinical Psychology Review, 31, 603-616. doi:
10.1016/j.cpr.2011.02.001
McLeod, B. D., & Weisz, J. R. (2005). The Therapy Process Observational Coding System Alliance scale: Measure characteristics and predition of outcome in usual clinical
practice. Journal of Consulting and Clinical Psychology, 73(2), 323-333. doi:
10.1037/0022-006X.73.2.323
McLeod, B. D., & Weisz, J. R. (2010). The Therapy process observational coding system for
child psychotherapy strategies scale. Journal of Clinical Child & Adolescent
Psychology, 39(3), 436-443. doi: 10.1080/15374411003691750

68

McPherson, P., Scribano, P., & Stevens, J. (2012). Barriers to successful treatment
completion in child sexual abuse survivors. Journal of Interpersonal Violence, 27(1),
23-39. doi: 10.1177/0886260511416466
Meiser-Stedman, R., Dalgleish, T., Glucksman, E., Yule, W., & Smith, P. (2009).
Maladaptive cognitive appraisals mediate the evolution of posttraumatic stress
reactions: A 6-month follow-up of child and adolescent assault and motor vehicle
accident survivors. Journal of Abnormal Psychology, 118(4), 778-787. doi:
10.1037/a0016945
Morris, A. S., Silk, J. S., Steinberg, L., Myers, S. M., & Robinson, L. R. (2007). The role of
the family context in the development of emotion regulation. Social Development,
16(2), 361-388. doi: 10.1111/j.1467-9507.2007.00389.x
Mossige, S., & Stefansen, K. (2007). Violence and sexual abuse against children and young
people in Norway: A questionnaire study among students in their last year of
secondary school. Oslo: Norwegian Social Research (NOVA).
Muran, J. C., Segal, Z. V., Samstag, L. W., & Crawford, C. E. (1994). Patient pretreatment
interpersonal problems and therapeutic alliance in short-term cognitive therapy.
Journal of Consulting and Clinical Psychology, 62(1), 185-190.
Muthn, L. K., & Muthn, B. O. (1998-2012). Mplus User's Guide. Seventh Edition. Los
Angeles, CA: Muthn & Muthn.
Nader, K., Newman, E., Weathers, F., Kaloupek, D. G., Kriegler, J., & Blake, D. (2004).
Clinician-Administered PTSD Scale for Children and Adolescents (CAPS-CA). Los
Angeles, CA: Western Psychological Press.
Nelson, T. D., Steele, R. G., & Mize, J. A. (2006). Practitioner attitudes
toward evidence-based practice: themes and challenges. Administration and Policy in Mental
Health and Mental Health Services Research, 33, 369-409. doi: 10.1007/s10488-0060044-4
NICE. (2005). Post-traumatic stress disorder (PTSD). The management of PTSD in adults and
children in primary and secondary care. from guidance.nice.org.uk/cg26
Nock, M. K., & Ferriter, C. (2005). Parent management of attendance and adherence in child
and adolescent therapy: A conceptual and empirical review. Clinical Child and Family
Psychology Review, 8(2), 149-166. doi: 10.1007/s10567-005-4753-0
Nock, M. K., & Kazdin, A. E. (2001). Parent expectancies for child therapy: assessment and
relation to participation in treatment. Journal of Child and Familiy Studies, 10(2),
155-180.
69

Norcross, J. C., & Lambert, M. J. (2011). Psychotherapy relationships that work II.
Psychotherapy, 48(1), 4-8. doi: 10.1037/a0022180
O'Callaghan, P., McMullen, J., Shannon, C., & Rafferty, H. (2013). A randomized controlled
trial of Trauma-Focused Cognitive Behavioral Therapy for sexually exploited, waraffected Congolese girls. Journal of the American Academy of Child and Adolesecent
Psychiatry, 52(4), 359-369. doi: 10.1016/j.jaac.2013.01.013
Orlinsky, D. E., Rnnestad, M. H., & Willutzki, U. (2004). Fifty years of psychotherapy
process-outcome research: continuity and change. In M. J. Lambert (Ed.), Bergin and
Garfield's Handbook of Psychotherapy and Behavior Change (Fifth ed., pp. 307-389).
New York: John Wiley & Sons, Inc.
Ormhaug, S. M., Jensen, T. K., Hukkelberg, S. S., Holt, T., & Egeland, K. (2012). Traumer
hos barn - blir de gjemt eller glemt? Kartlegging av traumatiske erfaringer hos barn og
unge henvist til BUP. Tidsskrift for norsk psykologforening, 49, 234-240.
Paivio, S. C., & Patterson, L. A. (1999). Alliance development in therapy for resolving child
abuse issues. Psychoterapy, 36, 343-354.
Pedhazur, E. J., & Schmelkin, L. P. (1991). Measurement, Design, and Analysis: An
Integrated Approach. Hillsdale, New Jersey: Lawrence Erbaum Associates.
Pinsof, W. M., & Catherall, D. R. (1986). The Integrative Psychotherapy Alliance: family,
couple and individual therapy scales. Journal of Marital and Family Therapy, 12(2),
137-151.
Pinsof, W. M., Zinbarg, R., & Knobloch-Fedders, L. M. (2008). Factorial and construct
validity of the revised short form integrative psychotherapy alliance scales for family,
couple and individual therapy. Family Process, 47(3), 281-301.
Powers, M. B., Halpern, J. M., Ferenschak, M. P., Gillihan, S. J., & Foa, E. B. (2010). A
meta-analytic review of prolonged exposure for posttraumatic stress disorder. Clinical
Psychology Review, 30, 635-641. doi: 10.1016/j.cpr.2010.04.007
Pynoos, R. S., Steinberg, A. M., & Piacentini, J. C. (1999). A developmental
psychopathology model of childhood traumatic stress and intersection with anxiety
disorders. Biological Psychiatry, 46(11), 1542-1554.
doi: http://dx.doi.org/10.1016/S0006-3223%2899%2900262-0
Reigstad, B., Jrgensen, K., & Wichstrm, L. (2006). Diagnosed and self-reported childhood
abuse in national and regional samples of child and adolescent psychiatric patients:
prevalence and correlates. Nordic Journal of Psychiatry, 60, 58-66. doi:
10.1080/08039480500504933
70

Ribbe, D. (1996). Psychometric review of Traumatic Event Screening Instrument for Children
(TESIC). In B. H. Stamm (Ed.), Measurement of stress, trauma, and adaptation (pp.
386-387). Lutherville, MD: Sidran Press.
Robbins, M. S., Liddle, H. A., Turner, C. W., Dakof, G. A., Alexander, J. F., & Kogan, S. M.
(2006). Adolescent and parent therapeutic alliances as predictors of dropout in
multidimensional family therapy. Journal of Family Psychology, 20(1), 108-116. doi:
10.1037/0893-3200.20.1.108
Robbins, M. S., Turner, C. W., Alexander, J. F., & Perez, G. A. (2003). Alliance and dropout
in family therapy for adolescents with behavior problems: individual and systemic
effects. Journal of Family Psychology, 17(4), 534-544. doi: 10.1037/08933200.17.4.534
Roos, J., & Werbart, A. (2013). Therapist and relationship factors influencing dropout from
individual psychotherapy: a literature review. Psychotherapy Research. doi:
DOI:10.1080/10503307.2013.775528
Rubin, D. B. (1976). Inference and missing data. Biometrika, 63(3), 581-592.
Safran, J. D., Crocker, P., McMain, S., & Murray, P. (1990). Therapeutic alliance rupture as a
therapy event for empirical investigation. Psychotherapy: Theory, Research, Practice,
Training, 27(2), 154-165. doi: 10.1037/0033-3204.27.2.154
Safran, J. D., & Muran, J. C. (2000). Negotiating the therapeutic alliance. A relational
treatment guide. New York: The Guilford Press.
Safran, J. D., & Muran, J. C. (2006). Has the concept of the therapeutic alliance outlived its
usefulness. Psychotherapy: Theory, Research, Practice, Training, 43(3), 286-291. doi:
10.1037/0033-3204.43.3.286
Saxe, G. N., Ellis, B. H., Fogler, J., & Navalta, C. P. (2012). Innovations in practice:
Preliminary evidence for effective family engagement in treatment for child traumatic
stress-trauma systems therapy approach to preventing dropout. Child and Adolescent
Mental Health, 17(1), 58-61. doi: 10.1111/j.1475-3588.2011.00626.x
Schafer, J. L., & Graham, J. W. (2002). Missing data: Our view of the state of the art.
Psychological Methods, 7(2), 147-177.
Scheeringa, M. S., Weems, C. F., Cohen, J. A., Amaya-Jackson, L., & Guthrie, D. (2011).
Trauma-focused cognitive-behavioral therapy for posttraumatic stress disorder in three
through six year-old children: A randomized clinical trial. The Journal of Child
Psychology and Psychiatry, 52(853-860). doi: 10.1111/j.1469-7610.2010.02354.x

71

Schou, L., Dyb, G., & Graff-Iversen , S. (2007). Voldsutsatt ungdom i Norge. Resultater fra
helseunderskelser i seks fylker (Vol. 8). Oslo: Norwegian Health Institute.
Schreiber, J. B., Stage, F. K., King, J., Nora, A., & Barlow, E. A. (2006). Reporting structural
equation modeling and confirmatory factor analyses: A review. The Journal of
Educational Research, 99(6), 323-337.
Shadish, W. R., Cook, T. D., & Campbell, D. T. (2002). Experimental and quasiexperimental designs for generalized causal inference. Boston, NY: Houghton Mifflin
Company.
Shelef, K., Diamond, G. M., Diamond, G. S., & Liddle, H. A. (2005). Adolescent and parent
alliance and treatment outcome in multidimensional family therapy. Journal of
Consulting and Clinical Psychology, 73(4), 689-698. doi: 10.1037/0022006X.73.4.689
Shirk, S. R., & Eltz, M. (1998). Multiple victimization and the process and outcome of child
psychotherapy. Journal of Aggression, Maltreatment & Trauma, 2(1), 233-251.
doi: http://dx.doi.org/10.1300/J146v02n01_13
Shirk, S. R., Gudmundsen, G., Kaplinski, H. C., & McMakin, D. L. (2008). Alliance and
outcome in cognitive-behavioral therapy for adolescent depression. Journal of Clinical
Child and Adolescent Psychology, 37(3), 631-639.
doi: http://dx.doi.org/10.1080/15374410802148061
Shirk, S. R., & Karver, M. S. (2010). Alliance in child and adolescent psychotherapy. In J. C.
Norcross (Ed.), Psychotherapy Relationships that Work: Evidence-Based
Responsiveness. New York: Oxford Press.
Shirk, S. R., Karver, M. S., & Brown, R. (2011). The alliance in child and adolescent
psychotherapy. Psychotherapy, 48(1), 17-24. doi: 10.1037/a0022181
Shirk, S. R., & Saiz, C. C. (1992). Clinical, empirical, and developmental perspectives on the
therapeutic relationship in child psychotherapy. Development and Psychopathology,
4(4), 713-728. doi: http://dx.doi.org/10.1017/S0954579400004946
Siegel, D. J. (2003). An interpersonal neurobiology of psychotherapy: The developing mind
and the resolution of trauma. In M. F. Solomon & D. J. Siegel (Eds.), Healing trauma:
Attachment, mind, body and brain (pp. 1-56). New York: Norton.
Silverman, W. K., Ortiz, C. D., & Viswesvaran, C. (2008). Evidence-based psychosocial
treatments for children and adolescents exposed to traumatic events. Journal of
Clinical Child & Adolescent Psychology, 37(1), 156-183. doi:
10.1080/15374410701818293
72

Sprang, G., Craig, C. D., Clark, J. J., Vergon, K., Tindall, M. S., Cohen, J. A., & Gurwitch, R.
(2013). Factors affecting the completion of trauma-focused treatments: What can
make a difference? Traumatology, 19(1), 28-40. doi: DOI:
10.1177/1534765612445931
Stallard, P., & Smith, E. (2007). Appraisals and cognitive coping styles associated with
chronic post-traumatic symptoms in road traffic accident survivors. Journal of Child
Psychology and Psychiatry, 48(2), 194-201. doi: 10.1111/j.1469-7610.2006.01692.x
Steinberg, A. M., Brymer, M. J., Decker, K. B., & Pynoos, R. S. (2004). The University of
California at Los Angeles post-traumatic stress disorder reaction index. Current
Psychiatry Reports, 6, 96-100.
Streiner, D. L. (2003). Starting at the beginning: An introduction to coefficient alpha and
internal consistency. Journal of Personality Assessment, 80, 99-103. doi:
10.1207/S15327752JPA8001_18
Sund, A. M., Larsson, B., & Wichstrm, L. (2001). Depressive symptoms among young
Norwegian adolescents as measured by The Mood and Feelings Questionnaire (MFQ).
European Child & Adolescent Psychiatry, 10(4), 222-229.
Tabachnick, B. G., & Fidell, L. S. (2007). Using multivariate statistics (Fifth ed.). Boston,
MA: Pearson Allyn & Bacon.
Tavakol, M., & Dennick, R. (2011). Making sense of Cronbach's alpha. International Journal
of Medical Education, 2, 53-55. doi: 10.5116/ijme.4dfb.8dfd
Thoresen, S., & Hjemdal, O. K. (2014). Vold og voldtekt i Norge. En nasjonal
forekomststudie av vold i et livslpsperspektiv (Vol. 1/2014): Nasjonalt
kunnskapssenter om vold og traumatisk stress (NKVTS).
Tolin, D. F., & Foa, E. B. (2008). Sex differences in trauma and posttraumatic stress disorder:
A quantitative review of 25 years of research. Psychological Trauma: Theory,
Research, Practice, and Policy, S(1), 37-85. doi: http://dx.doi.org/10.1037/19429681.S.1.37
Trickey, D., Siddaway, A. P., Meiser-Stedman, R., Serpell, L., & Field, A. P. (2012). A metaanalysis of risk factors for post-traumatic stress disorder in children and adolesecents.
Clinical Psychology Review, 32, 122-138. doi: 10.1016/j.cpr.2011.12.001
Tryon, G. S., Blackwell, S. C., & Hammel, E. F. (2007). A meta-analytic examination of
clienttherapist perspectives of the working alliance. Psychotherapy Research, 17(6),
629-642. doi: 10.1080/10503300701320611

73

Ulvenes, P. G., Berggraf, L., Hoffart, A., Stiles, T. C., Svartberg, M., McCullough, L., &
Wampold, B. E. (2012). Different processes for different therapies: therapist actions,
therapeutic bond, and outcome. Psychotherapy, 49(3), 291-302. doi:
10.1037/a0027895
Wampold, B. E., & Budge, S. L. (2012). The 2011 Leona Tyler Award Address: The
relationship - and its relationship to the common and specific factors of therapy. The
Counseling Psychologist, 40, 601-623. doi: 10.1177/0011000011432709
Wampold, B. E., & Imel, Z. E. (2015). General effects. Surving challenges and anticipating
additional evidence The Great Psychotherapy Debate. The Evidence for What Makes
Psychotherapy Work (2nd ed., pp. 178-212). New York, NY: Routledge.
Wampold, B. E., Imel, Z. E., Laska, K. M., Benish, S., Miller, M. E., Flckiger, C., . . .
Budge, S. (2010). Determining what works in the treatment of PTSD. Clinical
Psychology Review, 30, 923-933. doi: 10.1016/j.cpr.2010.06.005
Weinberger, J. (2014). Common factors are not som common and specific factors are not so
specified: Toward an inclusice integration of psychotherapy research. Psychotherapy,
51(4), 514-518. doi: 10.1037/a0037092
West, B. T. (2009). Analyzing longitudinal data with the linear mixed models procedure in
SPSS. Evaluation & the Health Professions, 32(3), 207-228. doi:
10.1177/0163278709338554
Zack, S. E., Castonguay, L. G., & Boswell, J. F. (2007). Youth working alliance: A core
clinical construct in need of empirical maturity. Harvard Review of Psychiatry, 15(6),
278-288. doi: 10.1080/10673220701803867

74

Appendices

Appendices
1a: CAPPATS, English version

CAPPATS
______________________
Patients Name

___________________
Date

1. I think my caregivers (e.g., mother/ father/ foster parent) like my therapist

1
Not at all

2
A Little

3
Most of the time

4
Very Much

2. I think my caregivers think it is important that I attend the sessions at the clinic
1
Not at all

2
A Little

3
Most of the time

4
Very Much

3. I think my caregivers want me to speak openly about my experiences to my therapist

1
Not at all

2
A Little

3
Most of the time

4
Very Much

4. I think my caregivers and I agree on what problems to work on

1
Not at all

2
A Little

3
Most of the time

4
Very Much

5. In think my caregivers think that the things the therapist and I do are helpful
1
Not at all

2
A Little

3
Most of the time

4
Very Much

Child- and Adolescent-Perceived Parental Approval of Treatment Scale

Jensen & Ormhaug, 2008

1b: CAPPATS, Norwegian version

CAPPATS

______________________
Navn

___________________
Dato

1. Jeg tror at mine foreldre (eks. mamma/ pappa/ fostermor/ fosterfar) liker terapeuten
min.
1
2
3
4
Stemmer ikke
Stemmer litt Stemmer for det meste Stemmer hele tiden
i det hele tatt

2. Jeg tror at foreldrene mine synes det er viktig at jeg kommer til timene her p BUP.
1
2
3
4
Stemmer ikke
Stemmer litt Stemmer for det meste Stemmer hele tiden
i det hele tatt

3. Jeg tror at foreldrene mine vil at jeg skal fortelle pent om det jeg har opplevd til
terapeuten min.
1
2
3
4
Stemmer ikke
Stemmer litt Stemmer for det meste Stemmer hele tiden
i det hele tatt

4. Jeg tror at foreldrene mine og jeg er enige i hvilke problemer som jeg skal jobbe med
terapeuten min for lse.
1
2
3
4
Stemmer ikke
Stemmer litt Stemmer for det meste Stemmer hele tiden
i det hele tatt

5. Jeg tror at foreldrene mine synes at det som terapeuten min og jeg gjr er hjelpsomt for
meg.
1
2
3
4
Stemmer ikke
Stemmer litt Stemmer for det meste Stemmer hele tiden
i det hele tatt

Child- and Adolescent-Perceived Parental Approval of Treatment Scale

Jensen & Ormhaug, 2008